Grand_Finale_Data_Report.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "Grand Finale Data Report.ipynb",
      "provenance": [],
      "collapsed_sections": [
        "2QH5SEwctitd"
      ]
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "F05KNxHOnFVh"
      },
      "source": [
        "# Location estimation using RSSI of wireless LAN\n",
        "\n",
        "---\n",
        "\n",
        "\n",
        "## Team C-LAMP\n",
        "\n",
        "### Team Members\n",
        "\n",
        "| Name                  | Designation               | Organisation                                                     |Email   |\n",
        "| -------------         | -------------             | -------------                                                    |--------|\n",
        "| Mr. Suresh V          | Associate Director        |CDAC Pune                                                         |sureshv@cdac.in|\n",
        "| Mr. Anil Kumar Gupta  | Associate Director        |CDAC Pune                                                         |anilg@cdac.in|\n",
        "| Ms. Irene S.          | Joint Director            |CDAC Chennai                                                      |irenes@cdac.in|\n",
        "| Mr. Pradeep Kumar CH  | Senior Technical Officer  |CDAC Pune                                                         |chpradeep@cdac.in|\n",
        "| Mr. Chirantan Ganguly | CDAC Research Intern      |Institute of Radio Physics and Electronics, University of Calcutta|chirantanganguly01@gmail.com|\n",
        "| Mr. Sagnik Nayak      | CDAC Research Intern      |Institute of Radio Physics and Electronics, University of Calcutta|sagniknayak2608@gmail.com|\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "gn5741sbnOuY"
      },
      "source": [
        "## Introduction\n",
        "\n",
        "In this project, our aim is to build a model to predict location of nodes based on RSSI readings from Access Points. The data for the project was provided by RISING, Japan for the problem ITU-ML5G-PS-016: Location estimation using RSSI of wireless LAN - organied as a part of the \"ITU AI/ML in 5G Challenge: applying machine learning in communication networks\". \n",
        "\n",
        "This project is divided into the following phases: \n",
        "\n",
        "- Phase 1: Emphasis is given on data preprocessing and exploration ,which will be covered in this report. Following this  the data and its attributes are described. Following this data preprocessing and transformation is done to efficiently fit the developed models. \n",
        "- Phase 2: Prediction model building is covered in this phase of the project."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "hhV8I_ROVwsW"
      },
      "source": [
        "##Background about RSSI\n",
        "\n",
        "- RSSI is ten times the logarithm of the ratio of power of the received signal and a reference power.\n",
        "\n",
        "- Power dissipates from a point source as it moves further out and the relationship between power and distance is that power is inversely proportional to the square of the distance travelled.\n",
        "\n",
        "- A direct theoretical relationship exist between  RSSI and distance:\n",
        "\n",
        "![picture](https://drive.google.com/uc?export=view&id=12XO_0vM6vPGa6agCAS1LZz9pKcncC2Qo)\n",
        "- With the distance of an unknown point known from a set of known points, its location can be estimated.\n",
        "\n",
        "- RSSI based technique is especially popular among researchers as it is readily available and no extra hardware is required."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "OHZ0GUn7XVq1"
      },
      "source": [
        "##Established Localization Techniques\n",
        "### Multilateration:\n",
        "\n",
        "If we can calculate the distance of a point from a set of 3 or more points whose location is known, we can localize the unknown point.\n",
        "\n",
        "In our scenario we have the following known parameters:\n",
        "- Location of the four APs (xi ,yi)\n",
        "- RSSI value measured at AP\n",
        "\n",
        "The following parameters are unknown and we must figure them out to undertake localization:\n",
        "- Distance of the RP from each AP (di)\n",
        "- Location of the RP (x,y)\n",
        "\n",
        "#### Distance measurement from RSSI:\n",
        "\n",
        "![picture](https://drive.google.com/uc?export=view&id=16ecbVE7V268iBYHac_lqeXCmqJ7h_PyV)\n",
        "\n",
        "As RSSI conveys information about strength of received signal, a relationship exists between RSSI and distance.\n",
        "\n",
        "![picture](https://drive.google.com/uc?export=view&id=1XYuJUX8YZ5vXUk4yKh9xPSBUcUGzR2Ml)\n",
        "\n",
        "n – Path loss Index (depends on Transmission environment)\n",
        "\n",
        "A - Average measured RSSI when the AP is 1 meter away from RP\n",
        "\n",
        "n, A can be found by proper experimentation in the environment.\n",
        "\n",
        "\n",
        "### RSSI Fingerprinting:\n",
        "\n",
        "In this method we directly estimate the location of the reference point without calculating its distance from the access point.\n",
        "\n",
        "#### Precondition for using Fingerprinting:\n",
        "Presence of large number of Reference Points in the \n",
        "training dataset, uniformly spread throughout the area.\n",
        "\n",
        "![picture](https://drive.google.com/uc?export=view&id=1NYhib1j62fNa1BX2s_KLrywh8pjZmg_u)\n",
        "\n",
        "#### Algorithms used for fingerprinting technique:\n",
        "  k-Nearest Neighbour.\n",
        "\n",
        "#### Advantages of using fingerprinting technique:\n",
        "High Accuracy in localization.\n",
        "\n",
        "#### Disadvantages of using fingerprinting technique:\n",
        "- Thorough Training phase required.\n",
        "- Sensitive to changes in Environment"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "nEH1lZBenWDf"
      },
      "source": [
        "##Dataset\n",
        "\n",
        "The provided data sets consist of the following files:\n",
        "\n",
        "1. AP_info.csv: CSV file providing the location (Latitude, Longitude) and environmental specification for a set of 4 access points (hereby referred to as APs).\n",
        "2. training_dataset.csv: CSV file providing us with the data acquired from four APs to be used for training the model. \n",
        "    \n",
        "    Data includes :\n",
        "    - TimeStamp(UNIX)\t\n",
        "    - Latitude\tof Reference Point\n",
        "    - Longitude\tof Reference Point\n",
        "    - SSID of Access Point\n",
        "    - Channel\t\n",
        "    - RSSI(dBm)\n",
        "\n",
        "  Data from 13 unique reference points are present in this dataset.\n",
        "\n",
        "3. verification_dataset.csv: CSV file providing us with the data acquired from the four APs to be used for testing the model. \n",
        "    \n",
        "    Data includes :\n",
        "    - TimeStamp(UNIX)\t\n",
        "    - Latitude\tof Reference Point\n",
        "    - Longitude\tof Reference Point\n",
        "    - SSID of Access Point\n",
        "    - Channel\t\n",
        "    - RSSI(dBm)\n",
        "\n",
        "  Data from 13 unique reference points are present in this dataset.\n",
        "\n",
        "The data in the datasets training_dataset.csv and  verification_dataset.csv have been collected from the same set of 4 APs at identical location.\n",
        "\n",
        "The data set also contains a timestamp on when the RSSI readings of the 4 APs was made. The each data set has 5200 observations.\n",
        "\n",
        "A map of the reference and the access points was provided with the dataset, as shown below:\n",
        "\n",
        "![picture](https://drive.google.com/uc?export=view&id=1UoF4iMbTNJ_323xPSADQPs_qTEKl5nSm)\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "TfP8SRWko9Kt"
      },
      "source": [
        "##Data Pre-processing\n",
        "\n",
        "### Preliminaries\n",
        "\n",
        "The following Python packages were used in this project.\n",
        "\n",
        "```\n",
        "import numpy\n",
        "import pandas \n",
        "import tensorflow \n",
        "from sklearn.preprocessing import MinMaxScaler\n",
        "from sklearn.model_selection import train_test_split\n",
        "from matplotlib import pyplot as plt\n",
        "```\n",
        "A short mentioning of the used packages:\n",
        "\n",
        "- numpy has been used for high-level mathematical functions\n",
        "- pandas has been mainly used for data manipulation and analysis.\n",
        "- tensorflow was used for training and inference of deep neural networks.\n",
        "- sklearn has many inbuilt functions containing various classification and regression algorithms which were used in this project\n",
        "- matplotlib was used in plotting various kinds of graphs and charts\n",
        "\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "KKQK3cVi1e0D",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "6f4c46ba-4add-4169-f4ec-0f4ba991252e"
      },
      "source": [
        "import numpy as np # linear algebra\n",
        "import pandas as pd # data processing, CSV file I/O (e.g. pd.read_csv)\n",
        "import tensorflow as tf\n",
        "from sklearn.preprocessing import MinMaxScaler\n",
        "from sklearn.model_selection import train_test_split\n",
        "from matplotlib import pyplot as plt\n",
        "import seaborn as sns\n",
        "import os\n",
        "import time\n",
        "cwd = os.getcwd()\n",
        "from google.colab import drive\n",
        "drive.mount('/content/drive')"
      ],
      "execution_count": 1,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Drive already mounted at /content/drive; to attempt to forcibly remount, call drive.mount(\"/content/drive\", force_remount=True).\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "6PFOmwTipQTg"
      },
      "source": [
        "Storing the path names"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "h-4RK3Ar2ALQ"
      },
      "source": [
        "AP_train_path='/content/drive/MyDrive/CDAC/RSSI Dataset Grand Finale/AP_info.csv'\n",
        "Wifi_train_path ='/content/drive/MyDrive/CDAC/RSSI Dataset Grand Finale/training_dataset.csv'\n",
        "Wifi_test_path ='/content/drive/MyDrive/CDAC/RSSI Dataset Grand Finale/verification_dataset.csv'"
      ],
      "execution_count": 2,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ZEQJnti1pTsW"
      },
      "source": [
        "Reading the csv files and storing in form of dataset"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Xv2QMkLfn5pt"
      },
      "source": [
        "AP_train=pd.read_csv(AP_train_path)"
      ],
      "execution_count": 3,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "HHNbEGvoph2M"
      },
      "source": [
        "Transforming the categorical data to numerical ones using one-hot-encoding"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "M1STa2Eqn_e9"
      },
      "source": [
        "AP_train = pd.concat([AP_train,pd.get_dummies(AP_train['Indoor/Outdoor'])['Outdoor']], axis=1, join='inner').drop(columns=['Indoor/Outdoor'])"
      ],
      "execution_count": 4,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "MrIFLJRmpZ6e"
      },
      "source": [
        "Let us have a look at the AP provided information"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "_ha74gr6n72F",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 175
        },
        "outputId": "414da985-e65d-417a-b9ff-4400cec34e79"
      },
      "source": [
        "AP_train"
      ],
      "execution_count": 5,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>AP SSID</th>\n",
              "      <th>Latitude</th>\n",
              "      <th>Longitude</th>\n",
              "      <th>Hight Difference(m)</th>\n",
              "      <th>Outdoor</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>1</td>\n",
              "      <td>35.157320</td>\n",
              "      <td>136.926447</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>2</td>\n",
              "      <td>35.157347</td>\n",
              "      <td>136.926807</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>3</td>\n",
              "      <td>35.157202</td>\n",
              "      <td>136.926827</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>4</td>\n",
              "      <td>35.157165</td>\n",
              "      <td>136.926470</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "   AP SSID   Latitude   Longitude  Hight Difference(m)  Outdoor\n",
              "0        1  35.157320  136.926447                    0        1\n",
              "1        2  35.157347  136.926807                    0        1\n",
              "2        3  35.157202  136.926827                    0        1\n",
              "3        4  35.157165  136.926470                    0        1"
            ]
          },
          "metadata": {},
          "execution_count": 5
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "BfJ_uzqpp1sE"
      },
      "source": [
        "AP_train = AP_train.astype({\"Latitude\":'float64',\"Longitude\":'float64',\"Hight Difference(m)\":'float64'})"
      ],
      "execution_count": 6,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "WrMoDDVOpsGT"
      },
      "source": [
        "Reading the measured data from the APs from training_dataset.csv and verification_dataset.csv and storing them inform of datasets."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "48C1sG81p409"
      },
      "source": [
        "train = pd.read_csv(Wifi_train_path)\n",
        "test = pd.read_csv(Wifi_test_path)"
      ],
      "execution_count": 7,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "engVZuFIqCzG"
      },
      "source": [
        "train.drop(columns=['No.'], inplace=True)\n",
        "test.drop(columns=['No.'], inplace=True)"
      ],
      "execution_count": 8,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "LGRDTj49p51k"
      },
      "source": [
        "Having a look at the datasets:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "XyW4Y_lnqDbm",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 424
        },
        "outputId": "214cb556-dcd9-4a72-82a1-0f2381d1f4ce"
      },
      "source": [
        "train"
      ],
      "execution_count": 9,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>TimeStamp(UNIX)</th>\n",
              "      <th>Latitude</th>\n",
              "      <th>Longitude</th>\n",
              "      <th>SSID</th>\n",
              "      <th>Channel</th>\n",
              "      <th>RSSI(dBm)</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>1631687696</td>\n",
              "      <td>35.157140</td>\n",
              "      <td>136.926306</td>\n",
              "      <td>1</td>\n",
              "      <td>11</td>\n",
              "      <td>-41</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>1631687697</td>\n",
              "      <td>35.157140</td>\n",
              "      <td>136.926306</td>\n",
              "      <td>1</td>\n",
              "      <td>11</td>\n",
              "      <td>-41</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>1631687697</td>\n",
              "      <td>35.157140</td>\n",
              "      <td>136.926306</td>\n",
              "      <td>1</td>\n",
              "      <td>11</td>\n",
              "      <td>-41</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>1631687697</td>\n",
              "      <td>35.157140</td>\n",
              "      <td>136.926306</td>\n",
              "      <td>1</td>\n",
              "      <td>11</td>\n",
              "      <td>-41</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>1631687697</td>\n",
              "      <td>35.157140</td>\n",
              "      <td>136.926306</td>\n",
              "      <td>1</td>\n",
              "      <td>11</td>\n",
              "      <td>-41</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>...</th>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>5195</th>\n",
              "      <td>1631688763</td>\n",
              "      <td>35.157236</td>\n",
              "      <td>136.926489</td>\n",
              "      <td>4</td>\n",
              "      <td>1</td>\n",
              "      <td>-52</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>5196</th>\n",
              "      <td>1631688763</td>\n",
              "      <td>35.157236</td>\n",
              "      <td>136.926489</td>\n",
              "      <td>4</td>\n",
              "      <td>1</td>\n",
              "      <td>-52</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>5197</th>\n",
              "      <td>1631688763</td>\n",
              "      <td>35.157236</td>\n",
              "      <td>136.926489</td>\n",
              "      <td>4</td>\n",
              "      <td>1</td>\n",
              "      <td>-52</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>5198</th>\n",
              "      <td>1631688763</td>\n",
              "      <td>35.157236</td>\n",
              "      <td>136.926489</td>\n",
              "      <td>4</td>\n",
              "      <td>1</td>\n",
              "      <td>-52</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>5199</th>\n",
              "      <td>1631688763</td>\n",
              "      <td>35.157236</td>\n",
              "      <td>136.926489</td>\n",
              "      <td>4</td>\n",
              "      <td>1</td>\n",
              "      <td>-52</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "<p>5200 rows × 6 columns</p>\n",
              "</div>"
            ],
            "text/plain": [
              "      TimeStamp(UNIX)   Latitude   Longitude  SSID  Channel  RSSI(dBm)\n",
              "0          1631687696  35.157140  136.926306     1       11        -41\n",
              "1          1631687697  35.157140  136.926306     1       11        -41\n",
              "2          1631687697  35.157140  136.926306     1       11        -41\n",
              "3          1631687697  35.157140  136.926306     1       11        -41\n",
              "4          1631687697  35.157140  136.926306     1       11        -41\n",
              "...               ...        ...         ...   ...      ...        ...\n",
              "5195       1631688763  35.157236  136.926489     4        1        -52\n",
              "5196       1631688763  35.157236  136.926489     4        1        -52\n",
              "5197       1631688763  35.157236  136.926489     4        1        -52\n",
              "5198       1631688763  35.157236  136.926489     4        1        -52\n",
              "5199       1631688763  35.157236  136.926489     4        1        -52\n",
              "\n",
              "[5200 rows x 6 columns]"
            ]
          },
          "metadata": {},
          "execution_count": 9
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "8Rd53y6RqFPL",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 424
        },
        "outputId": "f6b77ad6-aa3e-4a3e-de11-e894fe441f7c"
      },
      "source": [
        "test"
      ],
      "execution_count": 10,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>TimeStamp(UNIX)</th>\n",
              "      <th>Latitude</th>\n",
              "      <th>Longitude</th>\n",
              "      <th>SSID</th>\n",
              "      <th>Channel</th>\n",
              "      <th>RSSI(dBm)</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>1631689137</td>\n",
              "      <td>35.157261</td>\n",
              "      <td>136.926465</td>\n",
              "      <td>1</td>\n",
              "      <td>11</td>\n",
              "      <td>-41</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>1631689137</td>\n",
              "      <td>35.157261</td>\n",
              "      <td>136.926465</td>\n",
              "      <td>1</td>\n",
              "      <td>11</td>\n",
              "      <td>-41</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>1631689138</td>\n",
              "      <td>35.157261</td>\n",
              "      <td>136.926465</td>\n",
              "      <td>1</td>\n",
              "      <td>11</td>\n",
              "      <td>-41</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>1631689138</td>\n",
              "      <td>35.157261</td>\n",
              "      <td>136.926465</td>\n",
              "      <td>1</td>\n",
              "      <td>11</td>\n",
              "      <td>-41</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>1631689138</td>\n",
              "      <td>35.157261</td>\n",
              "      <td>136.926465</td>\n",
              "      <td>1</td>\n",
              "      <td>11</td>\n",
              "      <td>-41</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>...</th>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>5195</th>\n",
              "      <td>1631691291</td>\n",
              "      <td>35.157298</td>\n",
              "      <td>136.926575</td>\n",
              "      <td>4</td>\n",
              "      <td>1</td>\n",
              "      <td>-59</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>5196</th>\n",
              "      <td>1631691291</td>\n",
              "      <td>35.157298</td>\n",
              "      <td>136.926575</td>\n",
              "      <td>4</td>\n",
              "      <td>1</td>\n",
              "      <td>-59</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>5197</th>\n",
              "      <td>1631691291</td>\n",
              "      <td>35.157298</td>\n",
              "      <td>136.926575</td>\n",
              "      <td>4</td>\n",
              "      <td>1</td>\n",
              "      <td>-59</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>5198</th>\n",
              "      <td>1631691291</td>\n",
              "      <td>35.157298</td>\n",
              "      <td>136.926575</td>\n",
              "      <td>4</td>\n",
              "      <td>1</td>\n",
              "      <td>-59</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>5199</th>\n",
              "      <td>1631691292</td>\n",
              "      <td>35.157298</td>\n",
              "      <td>136.926575</td>\n",
              "      <td>4</td>\n",
              "      <td>1</td>\n",
              "      <td>-59</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "<p>5200 rows × 6 columns</p>\n",
              "</div>"
            ],
            "text/plain": [
              "      TimeStamp(UNIX)   Latitude   Longitude  SSID  Channel  RSSI(dBm)\n",
              "0          1631689137  35.157261  136.926465     1       11        -41\n",
              "1          1631689137  35.157261  136.926465     1       11        -41\n",
              "2          1631689138  35.157261  136.926465     1       11        -41\n",
              "3          1631689138  35.157261  136.926465     1       11        -41\n",
              "4          1631689138  35.157261  136.926465     1       11        -41\n",
              "...               ...        ...         ...   ...      ...        ...\n",
              "5195       1631691291  35.157298  136.926575     4        1        -59\n",
              "5196       1631691291  35.157298  136.926575     4        1        -59\n",
              "5197       1631691291  35.157298  136.926575     4        1        -59\n",
              "5198       1631691291  35.157298  136.926575     4        1        -59\n",
              "5199       1631691292  35.157298  136.926575     4        1        -59\n",
              "\n",
              "[5200 rows x 6 columns]"
            ]
          },
          "metadata": {},
          "execution_count": 10
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "jaiyuJulp9N5"
      },
      "source": [
        "Removing the duplicate data present :"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "qcQeHAjqqHDS"
      },
      "source": [
        "train.drop_duplicates(subset=['TimeStamp(UNIX)','Latitude','Longitude','SSID','RSSI(dBm)'],keep='first',inplace=True)\n",
        "test.drop_duplicates(subset=['TimeStamp(UNIX)','Latitude','Longitude','SSID','RSSI(dBm)'],keep='first',inplace=True)\n",
        "train.reset_index(inplace=True, drop=True)\n",
        "test.reset_index(inplace=True, drop=True)"
      ],
      "execution_count": 11,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "OYsitPtDqDvv"
      },
      "source": [
        "Merging the information of APs with the training and testing data:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "bkb72lS5qKKb"
      },
      "source": [
        "train_semi = train.merge(AP_train, how='inner', left_on='SSID', right_on='AP SSID').drop(columns=['AP SSID'])\n",
        "test_semi = test.merge(AP_train, how='inner', left_on='SSID', right_on='AP SSID').drop(columns=['AP SSID'])"
      ],
      "execution_count": 12,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "KF4s1SvhqKgE"
      },
      "source": [
        "Renaming the columns to make it more comprehensible"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "7YgyPSfoqOyT"
      },
      "source": [
        "train_semi.rename(columns = {'Latitude_x':'Latitude_OP','Longitude_x':'Longitude_OP', 'Latitude_y':'Latitude_AP', 'Longitude_y':'Longitude_AP'},inplace=True)\n",
        "test_semi.rename(columns = {'Latitude_x':'Latitude_OP','Longitude_x':'Longitude_OP', 'Latitude_y':'Latitude_AP', 'Longitude_y':'Longitude_AP'},inplace=True)"
      ],
      "execution_count": 13,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "SACIqu3wqOjY"
      },
      "source": [
        "As according to the research done in this domain, there is a clear cut relationship between RSSI and euclidean distance between the AP and the reference point (as shown in the formula below), hence we add a feature of the same.\n",
        "\n",
        "![picture](https://drive.google.com/uc?export=view&id=15tuh-QXcKEVx1P8EW9MuWRd_lJlTpywg)"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "dAG7ELMiqS2m"
      },
      "source": [
        "def euclidean_dist(lat_op, long_op, lat_ap, long_ap, height):\n",
        "  return (((lat_op - lat_ap)*111000)**2 + ((long_op - long_ap)*111000)**2 + height**2)**0.5"
      ],
      "execution_count": 14,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "aDpRDjVgqWEM"
      },
      "source": [
        "Since, we observe a significant variation in the provided RSSI data values (possibly due to multipath fading), we added a Moving Average of RSSI feature to smoothen the sudden variations in data."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "vsVobQr9qWPt"
      },
      "source": [
        "def moving_average(inp):\n",
        "  moving_avg=list(inp[\"RSSI(dBm)\"][:4])\n",
        "  i = 0\n",
        "  window_size=50\n",
        "  while i < len(list(inp[\"RSSI(dBm)\"])) - window_size + 1:\n",
        "      this_window = list(inp[\"RSSI(dBm)\"][i : i + window_size])\n",
        "      window_average = sum(this_window) / window_size\n",
        "      moving_avg.append(window_average)\n",
        "      i += 1\n",
        "  for z in range(len(list(inp[\"RSSI(dBm)\"]))-len(moving_avg)):\n",
        "    moving_avg.append(moving_avg[-1])\n",
        "  return moving_avg"
      ],
      "execution_count": 15,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Uk-md4Vhqb1O"
      },
      "source": [
        "train_semi['Distance'] = train_semi.apply(lambda row: euclidean_dist(row['Latitude_OP'],row['Longitude_OP'],row['Latitude_AP'],row['Longitude_AP'],row['Hight Difference(m)']),axis=1)\n",
        "test_semi['Distance'] = test_semi.apply(lambda row: euclidean_dist(row['Latitude_OP'],row['Longitude_OP'],row['Latitude_AP'],row['Longitude_AP'],row['Hight Difference(m)']),axis=1)"
      ],
      "execution_count": 16,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "FbFWnMgqqeYg"
      },
      "source": [
        "train_semi = pd.concat([train_semi,pd.DataFrame(moving_average(train_semi), columns = ['Moving Average (RSSI)'])], axis=1)\n",
        "test_semi = pd.concat([test_semi,pd.DataFrame(moving_average(test_semi), columns = ['Moving Average (RSSI)'])], axis=1)"
      ],
      "execution_count": 17,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "W6K_3FJDqj0Y"
      },
      "source": [
        "final_train = train_semi.copy()\n",
        "final_test = test_semi.copy()"
      ],
      "execution_count": 18,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "FJg81vLyqtNM",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 441
        },
        "outputId": "c5287755-92be-45a9-82af-dde70ea670db"
      },
      "source": [
        " train_semi.drop_duplicates(subset=['TimeStamp(UNIX)','Latitude_OP','Longitude_OP','RSSI(dBm)'])"
      ],
      "execution_count": 19,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>TimeStamp(UNIX)</th>\n",
              "      <th>Latitude_OP</th>\n",
              "      <th>Longitude_OP</th>\n",
              "      <th>SSID</th>\n",
              "      <th>Channel</th>\n",
              "      <th>RSSI(dBm)</th>\n",
              "      <th>Latitude_AP</th>\n",
              "      <th>Longitude_AP</th>\n",
              "      <th>Hight Difference(m)</th>\n",
              "      <th>Outdoor</th>\n",
              "      <th>Distance</th>\n",
              "      <th>Moving Average (RSSI)</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>1631687696</td>\n",
              "      <td>35.157140</td>\n",
              "      <td>136.926306</td>\n",
              "      <td>1</td>\n",
              "      <td>11</td>\n",
              "      <td>-41</td>\n",
              "      <td>35.157320</td>\n",
              "      <td>136.926447</td>\n",
              "      <td>0.0</td>\n",
              "      <td>1</td>\n",
              "      <td>25.439033</td>\n",
              "      <td>-41.00</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>1631687697</td>\n",
              "      <td>35.157140</td>\n",
              "      <td>136.926306</td>\n",
              "      <td>1</td>\n",
              "      <td>11</td>\n",
              "      <td>-41</td>\n",
              "      <td>35.157320</td>\n",
              "      <td>136.926447</td>\n",
              "      <td>0.0</td>\n",
              "      <td>1</td>\n",
              "      <td>25.439033</td>\n",
              "      <td>-41.00</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>1631687698</td>\n",
              "      <td>35.157140</td>\n",
              "      <td>136.926306</td>\n",
              "      <td>1</td>\n",
              "      <td>11</td>\n",
              "      <td>-41</td>\n",
              "      <td>35.157320</td>\n",
              "      <td>136.926447</td>\n",
              "      <td>0.0</td>\n",
              "      <td>1</td>\n",
              "      <td>25.439033</td>\n",
              "      <td>-41.00</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>1631687699</td>\n",
              "      <td>35.157140</td>\n",
              "      <td>136.926306</td>\n",
              "      <td>1</td>\n",
              "      <td>11</td>\n",
              "      <td>-41</td>\n",
              "      <td>35.157320</td>\n",
              "      <td>136.926447</td>\n",
              "      <td>0.0</td>\n",
              "      <td>1</td>\n",
              "      <td>25.439033</td>\n",
              "      <td>-41.00</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>1631687700</td>\n",
              "      <td>35.157140</td>\n",
              "      <td>136.926306</td>\n",
              "      <td>1</td>\n",
              "      <td>11</td>\n",
              "      <td>-41</td>\n",
              "      <td>35.157320</td>\n",
              "      <td>136.926447</td>\n",
              "      <td>0.0</td>\n",
              "      <td>1</td>\n",
              "      <td>25.439033</td>\n",
              "      <td>-42.16</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>...</th>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1071</th>\n",
              "      <td>1631688759</td>\n",
              "      <td>35.157236</td>\n",
              "      <td>136.926489</td>\n",
              "      <td>4</td>\n",
              "      <td>1</td>\n",
              "      <td>-52</td>\n",
              "      <td>35.157165</td>\n",
              "      <td>136.926470</td>\n",
              "      <td>0.0</td>\n",
              "      <td>1</td>\n",
              "      <td>8.152950</td>\n",
              "      <td>-58.24</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1072</th>\n",
              "      <td>1631688760</td>\n",
              "      <td>35.157236</td>\n",
              "      <td>136.926489</td>\n",
              "      <td>4</td>\n",
              "      <td>1</td>\n",
              "      <td>-52</td>\n",
              "      <td>35.157165</td>\n",
              "      <td>136.926470</td>\n",
              "      <td>0.0</td>\n",
              "      <td>1</td>\n",
              "      <td>8.152950</td>\n",
              "      <td>-58.24</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1073</th>\n",
              "      <td>1631688761</td>\n",
              "      <td>35.157236</td>\n",
              "      <td>136.926489</td>\n",
              "      <td>4</td>\n",
              "      <td>1</td>\n",
              "      <td>-52</td>\n",
              "      <td>35.157165</td>\n",
              "      <td>136.926470</td>\n",
              "      <td>0.0</td>\n",
              "      <td>1</td>\n",
              "      <td>8.152950</td>\n",
              "      <td>-58.24</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1074</th>\n",
              "      <td>1631688762</td>\n",
              "      <td>35.157236</td>\n",
              "      <td>136.926489</td>\n",
              "      <td>4</td>\n",
              "      <td>1</td>\n",
              "      <td>-52</td>\n",
              "      <td>35.157165</td>\n",
              "      <td>136.926470</td>\n",
              "      <td>0.0</td>\n",
              "      <td>1</td>\n",
              "      <td>8.152950</td>\n",
              "      <td>-58.24</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1075</th>\n",
              "      <td>1631688763</td>\n",
              "      <td>35.157236</td>\n",
              "      <td>136.926489</td>\n",
              "      <td>4</td>\n",
              "      <td>1</td>\n",
              "      <td>-52</td>\n",
              "      <td>35.157165</td>\n",
              "      <td>136.926470</td>\n",
              "      <td>0.0</td>\n",
              "      <td>1</td>\n",
              "      <td>8.152950</td>\n",
              "      <td>-58.24</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "<p>1059 rows × 12 columns</p>\n",
              "</div>"
            ],
            "text/plain": [
              "      TimeStamp(UNIX)  Latitude_OP  ...   Distance  Moving Average (RSSI)\n",
              "0          1631687696    35.157140  ...  25.439033                 -41.00\n",
              "1          1631687697    35.157140  ...  25.439033                 -41.00\n",
              "2          1631687698    35.157140  ...  25.439033                 -41.00\n",
              "3          1631687699    35.157140  ...  25.439033                 -41.00\n",
              "4          1631687700    35.157140  ...  25.439033                 -42.16\n",
              "...               ...          ...  ...        ...                    ...\n",
              "1071       1631688759    35.157236  ...   8.152950                 -58.24\n",
              "1072       1631688760    35.157236  ...   8.152950                 -58.24\n",
              "1073       1631688761    35.157236  ...   8.152950                 -58.24\n",
              "1074       1631688762    35.157236  ...   8.152950                 -58.24\n",
              "1075       1631688763    35.157236  ...   8.152950                 -58.24\n",
              "\n",
              "[1059 rows x 12 columns]"
            ]
          },
          "metadata": {},
          "execution_count": 19
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "PoeOGVgzvPxo"
      },
      "source": [
        "train[\"RSSI_AAA001\"] = np.where(train['SSID'] == 1, train[\"RSSI(dBm)\"], \"\")\n",
        "train[\"RSSI_AAA002\"] = np.where(train['SSID'] == 2, train[\"RSSI(dBm)\"], \"\")\n",
        "train[\"RSSI_AAA003\"] = np.where(train['SSID'] == 3, train[\"RSSI(dBm)\"], \"\")\n",
        "train[\"RSSI_AAA004\"] = np.where(train['SSID'] == 4, train[\"RSSI(dBm)\"], \"\")"
      ],
      "execution_count": 20,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "TvPz_1arvmdJ"
      },
      "source": [
        "train.drop(['SSID','RSSI(dBm)'],axis=1,inplace=True)"
      ],
      "execution_count": 21,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "BqGmrNzQFIeO"
      },
      "source": [
        "train_semi = pd.concat([train_semi,pd.get_dummies(train_semi['SSID'])], axis=1, join='inner').drop(columns=['SSID'])\n",
        "test_semi = pd.concat([test_semi,pd.get_dummies(test_semi['SSID'])], axis=1, join='inner').drop(columns=['SSID'])"
      ],
      "execution_count": 22,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "eeVlXAL9b0cc"
      },
      "source": [
        "Data restructuring:\n",
        "\n",
        "Data restructuring is an essential step of data preprocessing. \n",
        "We are changing the structure of the data provided to us by creating a  combination of AP and RP features, to suit model training."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "d2x9TR5xBE8l",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 458
        },
        "outputId": "b188a3ee-d7d0-4c82-97fa-e41755ab30c5"
      },
      "source": [
        "train_semi"
      ],
      "execution_count": 23,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>TimeStamp(UNIX)</th>\n",
              "      <th>Latitude_OP</th>\n",
              "      <th>Longitude_OP</th>\n",
              "      <th>Channel</th>\n",
              "      <th>RSSI(dBm)</th>\n",
              "      <th>Latitude_AP</th>\n",
              "      <th>Longitude_AP</th>\n",
              "      <th>Hight Difference(m)</th>\n",
              "      <th>Outdoor</th>\n",
              "      <th>Distance</th>\n",
              "      <th>Moving Average (RSSI)</th>\n",
              "      <th>1</th>\n",
              "      <th>2</th>\n",
              "      <th>3</th>\n",
              "      <th>4</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>1631687696</td>\n",
              "      <td>35.157140</td>\n",
              "      <td>136.926306</td>\n",
              "      <td>11</td>\n",
              "      <td>-41</td>\n",
              "      <td>35.157320</td>\n",
              "      <td>136.926447</td>\n",
              "      <td>0.0</td>\n",
              "      <td>1</td>\n",
              "      <td>25.439033</td>\n",
              "      <td>-41.00</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>1631687697</td>\n",
              "      <td>35.157140</td>\n",
              "      <td>136.926306</td>\n",
              "      <td>11</td>\n",
              "      <td>-41</td>\n",
              "      <td>35.157320</td>\n",
              "      <td>136.926447</td>\n",
              "      <td>0.0</td>\n",
              "      <td>1</td>\n",
              "      <td>25.439033</td>\n",
              "      <td>-41.00</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>1631687698</td>\n",
              "      <td>35.157140</td>\n",
              "      <td>136.926306</td>\n",
              "      <td>11</td>\n",
              "      <td>-41</td>\n",
              "      <td>35.157320</td>\n",
              "      <td>136.926447</td>\n",
              "      <td>0.0</td>\n",
              "      <td>1</td>\n",
              "      <td>25.439033</td>\n",
              "      <td>-41.00</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>1631687699</td>\n",
              "      <td>35.157140</td>\n",
              "      <td>136.926306</td>\n",
              "      <td>11</td>\n",
              "      <td>-41</td>\n",
              "      <td>35.157320</td>\n",
              "      <td>136.926447</td>\n",
              "      <td>0.0</td>\n",
              "      <td>1</td>\n",
              "      <td>25.439033</td>\n",
              "      <td>-41.00</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>1631687700</td>\n",
              "      <td>35.157140</td>\n",
              "      <td>136.926306</td>\n",
              "      <td>11</td>\n",
              "      <td>-41</td>\n",
              "      <td>35.157320</td>\n",
              "      <td>136.926447</td>\n",
              "      <td>0.0</td>\n",
              "      <td>1</td>\n",
              "      <td>25.439033</td>\n",
              "      <td>-42.16</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>...</th>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1071</th>\n",
              "      <td>1631688759</td>\n",
              "      <td>35.157236</td>\n",
              "      <td>136.926489</td>\n",
              "      <td>1</td>\n",
              "      <td>-52</td>\n",
              "      <td>35.157165</td>\n",
              "      <td>136.926470</td>\n",
              "      <td>0.0</td>\n",
              "      <td>1</td>\n",
              "      <td>8.152950</td>\n",
              "      <td>-58.24</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1072</th>\n",
              "      <td>1631688760</td>\n",
              "      <td>35.157236</td>\n",
              "      <td>136.926489</td>\n",
              "      <td>1</td>\n",
              "      <td>-52</td>\n",
              "      <td>35.157165</td>\n",
              "      <td>136.926470</td>\n",
              "      <td>0.0</td>\n",
              "      <td>1</td>\n",
              "      <td>8.152950</td>\n",
              "      <td>-58.24</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1073</th>\n",
              "      <td>1631688761</td>\n",
              "      <td>35.157236</td>\n",
              "      <td>136.926489</td>\n",
              "      <td>1</td>\n",
              "      <td>-52</td>\n",
              "      <td>35.157165</td>\n",
              "      <td>136.926470</td>\n",
              "      <td>0.0</td>\n",
              "      <td>1</td>\n",
              "      <td>8.152950</td>\n",
              "      <td>-58.24</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1074</th>\n",
              "      <td>1631688762</td>\n",
              "      <td>35.157236</td>\n",
              "      <td>136.926489</td>\n",
              "      <td>1</td>\n",
              "      <td>-52</td>\n",
              "      <td>35.157165</td>\n",
              "      <td>136.926470</td>\n",
              "      <td>0.0</td>\n",
              "      <td>1</td>\n",
              "      <td>8.152950</td>\n",
              "      <td>-58.24</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1075</th>\n",
              "      <td>1631688763</td>\n",
              "      <td>35.157236</td>\n",
              "      <td>136.926489</td>\n",
              "      <td>1</td>\n",
              "      <td>-52</td>\n",
              "      <td>35.157165</td>\n",
              "      <td>136.926470</td>\n",
              "      <td>0.0</td>\n",
              "      <td>1</td>\n",
              "      <td>8.152950</td>\n",
              "      <td>-58.24</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "<p>1076 rows × 15 columns</p>\n",
              "</div>"
            ],
            "text/plain": [
              "      TimeStamp(UNIX)  Latitude_OP  Longitude_OP  Channel  ...  1  2  3  4\n",
              "0          1631687696    35.157140    136.926306       11  ...  1  0  0  0\n",
              "1          1631687697    35.157140    136.926306       11  ...  1  0  0  0\n",
              "2          1631687698    35.157140    136.926306       11  ...  1  0  0  0\n",
              "3          1631687699    35.157140    136.926306       11  ...  1  0  0  0\n",
              "4          1631687700    35.157140    136.926306       11  ...  1  0  0  0\n",
              "...               ...          ...           ...      ...  ... .. .. .. ..\n",
              "1071       1631688759    35.157236    136.926489        1  ...  0  0  0  1\n",
              "1072       1631688760    35.157236    136.926489        1  ...  0  0  0  1\n",
              "1073       1631688761    35.157236    136.926489        1  ...  0  0  0  1\n",
              "1074       1631688762    35.157236    136.926489        1  ...  0  0  0  1\n",
              "1075       1631688763    35.157236    136.926489        1  ...  0  0  0  1\n",
              "\n",
              "[1076 rows x 15 columns]"
            ]
          },
          "metadata": {},
          "execution_count": 23
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "3AGg8gtXBGeA",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 458
        },
        "outputId": "45f8b65b-fc68-4f3a-af4f-15ceedbda353"
      },
      "source": [
        "test_semi"
      ],
      "execution_count": 24,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>TimeStamp(UNIX)</th>\n",
              "      <th>Latitude_OP</th>\n",
              "      <th>Longitude_OP</th>\n",
              "      <th>Channel</th>\n",
              "      <th>RSSI(dBm)</th>\n",
              "      <th>Latitude_AP</th>\n",
              "      <th>Longitude_AP</th>\n",
              "      <th>Hight Difference(m)</th>\n",
              "      <th>Outdoor</th>\n",
              "      <th>Distance</th>\n",
              "      <th>Moving Average (RSSI)</th>\n",
              "      <th>1</th>\n",
              "      <th>2</th>\n",
              "      <th>3</th>\n",
              "      <th>4</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>1631689137</td>\n",
              "      <td>35.157261</td>\n",
              "      <td>136.926465</td>\n",
              "      <td>11</td>\n",
              "      <td>-41</td>\n",
              "      <td>35.157320</td>\n",
              "      <td>136.926447</td>\n",
              "      <td>0.0</td>\n",
              "      <td>1</td>\n",
              "      <td>6.906728</td>\n",
              "      <td>-41.00</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>1631689138</td>\n",
              "      <td>35.157261</td>\n",
              "      <td>136.926465</td>\n",
              "      <td>11</td>\n",
              "      <td>-41</td>\n",
              "      <td>35.157320</td>\n",
              "      <td>136.926447</td>\n",
              "      <td>0.0</td>\n",
              "      <td>1</td>\n",
              "      <td>6.906728</td>\n",
              "      <td>-41.00</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>1631689139</td>\n",
              "      <td>35.157261</td>\n",
              "      <td>136.926465</td>\n",
              "      <td>11</td>\n",
              "      <td>-41</td>\n",
              "      <td>35.157320</td>\n",
              "      <td>136.926447</td>\n",
              "      <td>0.0</td>\n",
              "      <td>1</td>\n",
              "      <td>6.906728</td>\n",
              "      <td>-41.00</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>1631689140</td>\n",
              "      <td>35.157261</td>\n",
              "      <td>136.926465</td>\n",
              "      <td>11</td>\n",
              "      <td>-41</td>\n",
              "      <td>35.157320</td>\n",
              "      <td>136.926447</td>\n",
              "      <td>0.0</td>\n",
              "      <td>1</td>\n",
              "      <td>6.906728</td>\n",
              "      <td>-41.00</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>1631689141</td>\n",
              "      <td>35.157261</td>\n",
              "      <td>136.926465</td>\n",
              "      <td>11</td>\n",
              "      <td>-41</td>\n",
              "      <td>35.157320</td>\n",
              "      <td>136.926447</td>\n",
              "      <td>0.0</td>\n",
              "      <td>1</td>\n",
              "      <td>6.906728</td>\n",
              "      <td>-50.32</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>...</th>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1096</th>\n",
              "      <td>1631691288</td>\n",
              "      <td>35.157298</td>\n",
              "      <td>136.926575</td>\n",
              "      <td>1</td>\n",
              "      <td>-59</td>\n",
              "      <td>35.157165</td>\n",
              "      <td>136.926470</td>\n",
              "      <td>0.0</td>\n",
              "      <td>1</td>\n",
              "      <td>18.773415</td>\n",
              "      <td>-69.32</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1097</th>\n",
              "      <td>1631691289</td>\n",
              "      <td>35.157298</td>\n",
              "      <td>136.926575</td>\n",
              "      <td>1</td>\n",
              "      <td>-59</td>\n",
              "      <td>35.157165</td>\n",
              "      <td>136.926470</td>\n",
              "      <td>0.0</td>\n",
              "      <td>1</td>\n",
              "      <td>18.773415</td>\n",
              "      <td>-69.32</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1098</th>\n",
              "      <td>1631691290</td>\n",
              "      <td>35.157298</td>\n",
              "      <td>136.926575</td>\n",
              "      <td>1</td>\n",
              "      <td>-59</td>\n",
              "      <td>35.157165</td>\n",
              "      <td>136.926470</td>\n",
              "      <td>0.0</td>\n",
              "      <td>1</td>\n",
              "      <td>18.773415</td>\n",
              "      <td>-69.32</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1099</th>\n",
              "      <td>1631691291</td>\n",
              "      <td>35.157298</td>\n",
              "      <td>136.926575</td>\n",
              "      <td>1</td>\n",
              "      <td>-59</td>\n",
              "      <td>35.157165</td>\n",
              "      <td>136.926470</td>\n",
              "      <td>0.0</td>\n",
              "      <td>1</td>\n",
              "      <td>18.773415</td>\n",
              "      <td>-69.32</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1100</th>\n",
              "      <td>1631691292</td>\n",
              "      <td>35.157298</td>\n",
              "      <td>136.926575</td>\n",
              "      <td>1</td>\n",
              "      <td>-59</td>\n",
              "      <td>35.157165</td>\n",
              "      <td>136.926470</td>\n",
              "      <td>0.0</td>\n",
              "      <td>1</td>\n",
              "      <td>18.773415</td>\n",
              "      <td>-69.32</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "<p>1101 rows × 15 columns</p>\n",
              "</div>"
            ],
            "text/plain": [
              "      TimeStamp(UNIX)  Latitude_OP  Longitude_OP  Channel  ...  1  2  3  4\n",
              "0          1631689137    35.157261    136.926465       11  ...  1  0  0  0\n",
              "1          1631689138    35.157261    136.926465       11  ...  1  0  0  0\n",
              "2          1631689139    35.157261    136.926465       11  ...  1  0  0  0\n",
              "3          1631689140    35.157261    136.926465       11  ...  1  0  0  0\n",
              "4          1631689141    35.157261    136.926465       11  ...  1  0  0  0\n",
              "...               ...          ...           ...      ...  ... .. .. .. ..\n",
              "1096       1631691288    35.157298    136.926575        1  ...  0  0  0  1\n",
              "1097       1631691289    35.157298    136.926575        1  ...  0  0  0  1\n",
              "1098       1631691290    35.157298    136.926575        1  ...  0  0  0  1\n",
              "1099       1631691291    35.157298    136.926575        1  ...  0  0  0  1\n",
              "1100       1631691292    35.157298    136.926575        1  ...  0  0  0  1\n",
              "\n",
              "[1101 rows x 15 columns]"
            ]
          },
          "metadata": {},
          "execution_count": 24
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "vifHsPb9lcej"
      },
      "source": [
        "Data Smoothening:\n",
        "\n",
        "Smoothening of the data is required as the RSSI values tend to change abruptly over time. \n",
        "\n",
        "Various techniques for smoothening exists - in our case we are employing the simple moving averaging technique of RSSI values to undertake data smoothening."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "9SpwPKdHCkyc",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 295
        },
        "outputId": "d52a9b19-5fa5-41ec-8716-a639797437a6"
      },
      "source": [
        "lRSSI=list(train_semi['RSSI(dBm)'])\n",
        "lMoving_avg=list(train_semi['Moving Average (RSSI)'])\n",
        "plt.plot(lRSSI, color=\"blue\", label=\"Original Data\")\n",
        "plt.plot(lMoving_avg, color=\"orange\", label=\"Smoothened Data\")\n",
        "plt.title(\"Data Smoothening\")\n",
        "plt.xlabel(\"No. of Observations\")\n",
        "plt.ylabel(\"RSSI (dBm)\")\n",
        "plt.legend(loc='best',fancybox=True, shadow=True)\n",
        "plt.show()"
      ],
      "execution_count": 25,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "2TJtjfh5lgo3"
      },
      "source": [
        "Visualising the Data Points:\n",
        "\n",
        "Plotting the Reference Points and the Access Points given in the dataset."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "UgTNJCae2GbC",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 295
        },
        "outputId": "b4751e12-f301-49b9-a5d0-46c4c5642c6a"
      },
      "source": [
        "lat_RP = set(test_semi['Latitude_OP'])\n",
        "long_RP = set(test_semi['Longitude_OP'])\n",
        "lat_AP = set(test_semi['Latitude_AP'])\n",
        "long_AP = set(test_semi['Longitude_AP'])\n",
        "plt.scatter(x=list(long_RP),y=list(lat_RP), color=\"blue\", label=\"Reference Points\")\n",
        "plt.scatter(x=list(long_AP),y=list(lat_AP), color=\"red\", label=\"Access Points\")\n",
        "plt.title(\"Visualising Validation Dataset\")\n",
        "plt.xlabel(\"Longitude\")\n",
        "plt.ylabel(\"Latitude\")\n",
        "plt.legend(loc='best',fancybox=True, shadow=True)\n",
        "plt.show()"
      ],
      "execution_count": 26,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAZsAAAEWCAYAAACwtjr+AAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nO3deZwV1Z338c8XcAE3XDrGjcVdNhFblCQuxChoEplEHHFIFKOSRU2iUaNjJhojT/RhRo1GozzBvQdEooZJMDpukcSNBsEdgwgoagKICKLI8nv+qNNwaW/fvkBXb3zfr9d99b2nTp06VRfur+qcU6cUEZiZmeWpTVNXwMzMWj8HGzMzy52DjZmZ5c7BxszMcudgY2ZmuXOwMTOz3DnYWKOS9LKko3LeRkjaO72/WdJ/NId6rS9Jl0u6O73vJGmppLb15d3AbTW7/bfWxcHGGoykP0u6okj6IEnvSWoXEd0j4onGqlNEfC8ifllGvgavl6TDJH0kaesiy56XdE65ZUXE3IjYOiJWNUC9bpd0Za3yc/leJD0h6RNJSyR9KGmKpIslbbEeZaw5echTY21nU+VgYw3pDuBbklQr/dtAVUSsbII6NZmIeAZ4GxhcmC6pB9ANGNMU9WoC50TENsAuwE+AIcDEIv9OrBVzsLGG9ACwI3B4TYKk7YGvAXemz7MlfSW97yupOp3x/kPSNSn9KElvFxZcZL2nJX0g6V1Jv5G0ebEKFZ7FS9pJ0h/Teu9LmiSpTZHyL5c0TtKd6Yz8ZUmVBWX2SVcmSyTdK+me2lcKBe4ATq2VdiowMSIWSvq1pLcKzvoPL1IGkrqkM+926XNXSX9JdfhfYKda+e9NV5OLJT0pqXtKHw4MBS5KzXL/U2T/t5B0naR30uu6miuRmu9G0k8k/TMd/9Pr2Pd1RMRH6erpBKAf8NVUZp3fp6Qn0+rTU31PlrR9+h7nS1qU3u9esO/DJM1Kx+ZNSUMLln1H0qtpvYckda5rO+Xsk5XPwcYaTER8DIxj3R/XfwVei4jpRVb5NfDriNgW2CutW45VwHlkP7D9gKOBH5Sx3k/IrjQqgJ2Bfwfqmq/pBGAs0BGYAPwGIP0I3g/cDuxAdnXyjRLbvAs4QtIeaf02wL+RBSGAyUDvVNZ/A/dK2rKMfflvYArZMfglcFqt5Q8C+wCfA6YCVQARMSq9/7+pWe7rRcq+FDgs1etAoC/ws4Llnwe2A3YDzgBuTCcVZYmIuUA1a09K6vw+I+KIlOfAVN97yH63bgM6A52Aj1n7/WwFXA8cl66mvgBMS8sGkX3n3yT7NzCJdHVZx3asATnYbABJv5T0gqRpkh6WtGsd+ValPNMkTShIP0fSzHSmulNB+lHpTLRmnZ+n9P0K0qals+Af11PHIyRNlbRS0uBSeRvYHcDggh/MU1n7w1rbCmBvSTtFxNLU7FSviJgSEc9ExMqImA3cAhxZxqoryJpyOkfEioiYFHVPDvjXiJiY+kjuIvvRhexHuB1wfSrjPuC5EnV9C3iCrCkRsh/SLYA/peV3R8TCtC//lZbtV2onJHUCDgH+IyKWR8STwP/U2u6tEbEkIpYDlwMHStquVLkFhgJXRMQ/I2I+8IuC+kN2HK9I+z8RWFpfnYt4hyzArvf3mY7X7yNiWUQsAUbUyr8a6CGpfUS8GxEvp/TvAb+KiFdTk+7/AXrXXN1Yvhxs6pECwO21kkdGRK+I6A38Efh5Hat/HBG90+uEgvS/AV8B5hRZZ1LBOlcARMSMmjTgYGAZ2dl1KXOBYWRnwI0mIv4KLAD+RdJeZGfFddXhDGBf4DVJkyV9rZxtSNo3NZ28J+lDsh+NnepbDxgJzAQeTs0sF5fI+17B+2XAlqkJa1dgXq0g9VY9272DtT/W3wbGRsSKtC8XpGadxZI+ILtiqG9fdgUWRcRHBWlr/i1JaivpKklvpOMzOy0q5xjVlF/4b3NOSquxsFb/2zLgM4Mg6rEb8H6q73p9n5I6SLpF0pyU/0mgo6S26ZicTBZY3pX0J0n7p1U7A79OzXUfpO0r1cVy5mCzASLiw4KPW1F3U0xd6z+fzuA2xNHAGxExB0DSXspGgU1R1gexf9rG7Ih4gewsr7HdSXZF8y3goYj4R7FMEfH3iDiFrKnnamB8agb5COhQk0/ZcN+KglV/C7wG7JOa4P6d7EejpHSm/5OI2JOsmex8SUev5769C+wmrdO5vUc969wH7C6pP1kTzh0AyvpnLiJratw+IjoCi8vYl3eB7dOxqtGp4P2/AYPITmi2A7qk9Jpy6/v3+g7ZD3Nh2e/Us07ZUpPiwWTNWLD+3+dPyK6kDk35a5rABBARD0XEMWRXsa8B/y8tfwv4bkR0LHi1j4inGmrfrG4ONhtI0ghJb5E1OdR1ZbOlsg7wZyT9S5lF95M0XdKDSp26tQxh3VFMo4BzI+Jg4ALgpnL3IUd3kv3QnUXdTWhI+pakiohYDXyQklcDr5Mdu69K2oysv6BwqOw2wIfA0hRcv19OpSR9TdLeKVAsJusrWN9g/HRa7xxJ7VI/QN9SK6Sz7fFk/QxzIqK6YD9WAvOBdqnZdNv6KpBONKqBX0jaXNKXgMK+l22A5cBCsqD9f2oV8Q9gzxKbGAP8TFJFaub9ObDB9/DUSFckRwJ/IGt6nFhQ31LfZ+36bkPWT/OBpB2Aywq2sbOyofZbkR2Dpaz9jm8GLtHawRLbSTqpxHasATnY1EHSs5KmAb8DTijoLxkAEBGXRsQeZJ2tdd0v0TkiKsnONK9LzUqlTE3rHAjcQDa6q7BOm5Odkd+bPm9N1gF6b6rrLWRnc00qXbU9RXbVN6FE1oHAy5KWkg0WGBIRH0fEYrIO4t8B88iudApHp11AdkyXkJ21ltuZuw/wCNkP0NPATRHxeJnrAhARn5JdnZxBFiC/RdaUuryeVe8gu1q4syDtIeDPZMF1DvAJ9TfJ1fg34FCypqDLapV7ZypvHvAKULsvbDTQLTUnPcBnXUkWzF4AXiT7d1nXaLty/EbSErIf8+uA3wMD00kG1P99Xg7cker7r6mM9mTNtc+QHcMabYDzya7E3ifry/k+QETcT3YFPTY1v70EHFdiO9aAVHf/qEHWZwMMi4hhdSzvRDaMtUc95dwO/DEixhekzQYqI2JBHeusszydRZ8dEcemz9sCMyKizgBTbLvWsCQ9C9wcEbc1dV3Mmitf2WwASfsUfBxE1i5cO8/2Wntvwk7AF8nOMkuV+/mavgBJfcm+n4UFWU6hoAkt9R29WdMUoMyBWK4kHZm+q3aSTgN6se7ZtZnV4mCzYa6S9JKkF4BjgR8BSKqU9LuU5wCgWtJ04HHgqoh4JeX7obKbFncHXihYZzDwUlrnerJmpUjrbAUcQ9bZXGgocEZa52Wy4IekQ9I2TgJukfQy1lD2A6aTNaP9BBgcEe82bZXMmjc3o5mZWe58ZWNmZrlr19QVaI522mmn6NKlS1NXw8ysRZkyZcqCiKgotszBpoguXbpQXV1df0YzM1tDUrFZUQA3o5mZWSNwsDEzs9w52JiZWe7cZ2Nmufn000954403WLZsWVNXxRpQhw4d2Guvvdh886LPLCzKwcbMcvPGG2/QsWNH9ttvP9q0cUNKa7B69Wr+8Y9/MHPmTLp161b2ev72m1hVFXTpAm3aZH+rqpq6RmYNZ9myZey8884ONK1ImzZt2HnnnVm2bBlPPfUU5U4M4H8BTaiqCoYPhzlzICL7O3y4A461Lg40rU+bNm2QxDPPPMOcOXWOdl53nZzrZCVceinUbspetixLNzNr7iSxZMmSsvI62DShuXPXL93M1l/btm3p3bs3PXr04Otf/zoffPBByfzz58/n0EMP5aCDDmLSpEkl8zaVYcOG0bVrV3r37k2fPn14+umnS+b/whe+UG+Z1113Xa4DORxsmlCnTuuXbmbrr3379kybNo2XXnqJHXbYgRtvvLFk/kcffZSePXvy/PPPc/jhh5e1jVWrVjVEVdfLyJEjmTZtGldddRXf/e53S+Z96qn6n3ztYNOKjRgBHTqsm9ahQ5ZutinKe8BMv379mDdvHpCNlBs4cCAHH3wwhx9+OK+99hrTpk3joosu4g9/+AO9e/fm448/5uGHH6Zfv3706dOHk046iaVLlwLZtFY//elP6dOnD/fee2/JfJdddhl9+vShZ8+evPZa9virpUuXcvrpp9OzZ0969erF73//e4A6y6nLEUccwcyZMwG45ppr6NGjBz169OC6665bk2frrbcG4IknnuCoo45i8ODB7L///gwdOpSI4Prrr+edd96hf//+9O/fn1WrVjFs2DB69OhBz549ufbaazf+4EeEX7VeBx98cDSWu++O6Nw5Qsr+3n13o23aLHfV1dVl57377ogOHSKy4TLZq0OHjf8/sdVWW0VExMqVK2Pw4MHx4IMPRkTEl7/85Xj99dcjIuKZZ56J/v37R0TEbbfdFmeffXZERMyfPz8OP/zwWLp0aUREXHXVVfGLX/wiIiI6d+4cV199dVn5rr/++oiIuPHGG+OMM86IiIiLLroofvSjH62p5/vvv1+ynEKnnXZa3HvvvRERMW7cuOjbt29UV1dHjx49YunSpbFkyZLo1q1bTJ06dZ1j8Pjjj8e2224bb731VqxatSoOO+ywmDRp0pp6zp8/PyKy7+0rX/nKmu0tWrToM3Worq6Oa665Jl544YU1aUB11PG76vtsmtjQodnLbFNXasDMxvwf+fjjj+nduzfz5s3jgAMO4JhjjmHp0qU89dRTnHTSSWvyLV++/DPrPvPMM7zyyit88YtfBLKbVPv167dm+cknn1xWvm9+85sAHHzwwdx3X/b8w0ceeYSxY8euybP99tvzxz/+sWQ5hS688EKuvPJKKioqGD16NI8++ijf+MY32GqrrdZsc9KkSRx00EHrrNe3b1923313AHr37s3s2bP50pe+tE6ePffck1mzZnHuuefy1a9+lWOPPbb4wV0PDjZm1izkNWCmps9m2bJlDBgwgBtvvJFhw4bRsWNHpk2bVnLdiOCYY45hzJgxRZfX/LDXl2+LLbYAssEKK1eu3ODtFRo5ciSDBw9e8/nRRx+td53CupSqz/bbb8/06dN56KGHuPnmmxk3bhy33nprWeXXxX02LZhvCLXWJO8BMx06dOD666/nv/7rv+jQoQNdu3bl3nvvBbIf+enTp39mncMOO4y//e1va/pEPvroI15//fUNzlfomGOOWWewwqJFizaonBqHH344DzzwAMuWLeOjjz7i/vvvL3uAA8A222yzZhjzggULWL16NSeeeCJXXnklU6dOLbucujjYtFC+IdRam8YYMHPQQQfRq1cvxowZQ1VVFaNHj+bAAw+ke/fu/OEPf/hM/oqKCm6//XZOOeUUevXqRb9+/dZ08G9IvkI/+9nPWLRoET169ODAAw/k8ccf36ByavTp04dhw4bRt29fDj30UM4888zPNKGVMnz4cAYOHEj//v2ZN28eRx11FL179+Zb3/oWv/rVr8oupy6KMqca2JRUVlZGc394WpcuWYCprXNnmD27sWtjVtyUKVM4+OCDy85fVZX10cydm13RjBjhPs3masqUKTz55JN85StfoWfPngBImhIRlcXyu8+mhfINodYaecBM6+VmtBbKN4SaWUviYNNC+YZQM2tJcg02kgZKmiFppqSLiyzfQtI9afmzkroULLskpc+QNGA9yrxe0tKCz8MkzZc0Lb3ObPg9bXxDh8KoUVkfjZT9HTXKTRBm1jzl1mcjqS1wI3AM8DYwWdKEiHilINsZwKKI2FvSEOBq4GRJ3YAhQHdgV+ARSfumdeosU1IlsH2R6twTEec0/F42Lbdvm1lLkeeVTV9gZkTMiohPgbHAoFp5BgF3pPfjgaMlKaWPjYjlEfEmMDOVV2eZKbiNBC7KcZ/MzGwD5BlsdgPeKvj8dkormiciVgKLgR1LrFuqzHOACRHxbpG6nCjpBUnjJe1RrLKShkuqllQ9f/78cvbPzFqIBx54AEll37OSt8svv5zddtttzaMPJkyYUDL/8ccfX++jEW6//Xbeeeedhqxmg2oVAwQk7QqcBNxQZPH/AF0iohfwv6y9klpHRIyKiMqIqKyoqMivsmbW6MaMGcOXvvSlsqaBaSznnXce06ZN49577+U73/kOq1evrjPvxIkT6dixY8nyNuVgMw8ovIrYPaUVzSOpHbAdsLDEunWlHwTsDcyUNBvoIGkmQEQsjIiaGfZ+B5R/h5mZNa4c5mBaunQpf/3rXxk9evQ6E1+uWrWKCy64gB49etCrVy9uuCE7V508eTJf+MIXOPDAA+nbty9Llixh1apVXHjhhRxyyCH06tWLW265BYB3332XI444Ys0VyqRJk9Z7ev4DDjiAdu3asWDBAsaMGUPPnj3p0aMHP/3pT9fk6dKlCwsWLGD27NkccMABnHXWWXTv3p1jjz2Wjz/+mPHjx1NdXc3QoUPXPBrh4osvplu3bvTq1YsLLrhgo4/jRqtrOuiNfZENPpgFdAU2B6YD3WvlORu4Ob0fAoxL77un/Fuk9WcBbcspM62/tOD9LgXvvwE8U1/dG/MRA2at2fo8YiCvZwzcfffd8Z3vfCciIvr167emTjfddFOceOKJsWLFioiIWLhwYSxfvjy6du0azz33XERELF68OFasWBG33HJL/PKXv4yIiE8++SQOPvjgmDVrVvznf/5nXHnllRGRPcLgww8/LGt6/ssuuyxGjhwZEdnjDXbZZZd4++23Y4899oh//vOfsWLFiujfv3/cf//9EbF2+v8333wz2rZtG88//3xERJx00klx1113RUTEkUceGZMnT46IiAULFsS+++4bq1evrrMOG2t9HzGQ25VNZH0w5wAPAa+mQPKypCsknZCyjQZ2TFch5wMXp3VfBsYBrwB/Bs6OiFV1lVlPVX4o6WVJ04EfAsMacj/NrIGUesbARhgzZgxDhgwBYMiQIWua0h555BG++93v0q5dNih3hx12YMaMGeyyyy4ccsghAGy77ba0a9eOhx9+mDvvvJPevXtz6KGHsnDhQv7+979zyCGHcNttt3H55Zfz4osvss0226wzPf+f//xntt1226L1uvbaa+nduzcXXHAB99xzD9XV1Rx11FFUVFTQrl07hg4dypNPPvmZ9WoeBw3ZIwtmF5mfarvttmPLLbfkjDPO4L777qND7ZvymkCu09VExERgYq20nxe8/4Ssr6XYuiOAz9yiWKzMInm2Lnh/CXDJelXczBpfDnMwvf/++zz22GO8+OKLSGLVqlVIYuTIketVTkRwww03MGDAgM8se/LJJ/nTn/7EsGHDOP/88zn11FPLmp7/vPPOW6d5q9hEoMXUfkTAxx9//Jk87dq147nnnuPRRx9l/Pjx/OY3v+Gxxx4rq/y8tIoBAmbWCuQwB9P48eP59re/zZw5c5g9ezZvvfUWXbt2ZdKkSRxzzDHccssta57n8v7777Pffvvx7rvvMnnyZACWLFnCypUrGTBgAL/97W9ZsWIFAK+//jofffQRc+bMYeedd+ass87izDPPZOrUqRs8PX/fvn35y1/+woIFC1i1ahVjxozhyCOPLHtfCx8RsHTpUhYvXszxxx/PtddeW/TxCY3NE3GaWfMwYkT2nIzCprSNnINpzJgx63S0A5x44omMGTOGG264gddff51evXqx2WabcdZZZ3HOOedwzz33cO655/Lxxx/Tvn17HnnkEc4880xmz55Nnz59iAgqKip44IEHeOKJJxg5ciSbbbYZW2+9NXfeeSfz5s3j9NNPXzO6rNzp+XfZZReuuuoq+vfvT0Tw1a9+lUGDat+aWLdhw4bxve99j/bt2/Pggw8yaNAgPvnkEyKCa665pvyDlhM/YqCIlvCIAbOWYH0fMeBnDLQcfsSAmbVcnoOp1XKfjZmZ5c7BxsxyVerOeGuZNuQ7dbAxs9x06NCBd9991wGnFVm9ejXvvffempF55XKfjVkL0FL7zffaay+mTp3Ke++919RVsQa0YsUK5s6diyTatCnvmsXBxqyZq6pad0TwnDnZZ2j+AWfzzTdn5cqVPPXUU3Ts2JHsCSLWGnz66ae0adOGz33uc2Xl99DnIjz02ZqTLl2yAFNb585QZKaSZmfVqlU888wzzJo1i1WrVjV1dayBtG/fnsMOO4zOnTuvSSs19NnBpggHG2tO2rTJZqWsTQJ3hVhzUirYeICAWTOXwywuZo3OwcasmRsxIpu1pdBGzuJi1ugcbMyauaFDYdSorI9Gyv6OGtX8BweYFfJoNLMWwLO4WEvnKxszM8udg42ZmeXOwcbMzHLnYGNmZrlzsDEzs9w52JiZWe4cbMzMLHcONmZmlrtcg42kgZJmSJop6eIiy7eQdE9a/qykLgXLLknpMyQNWI8yr5e0tJxtmJlZ48gt2EhqC9wIHAd0A06R1K1WtjOARRGxN3AtcHVatxswBOgODARuktS2vjIlVQLbl7MNaxmqqrIp9tu0yf5WVTV1jcxsQ+R5ZdMXmBkRsyLiU2AsMKhWnkHAHen9eOBoZU9XGgSMjYjlEfEmMDOVV2eZKRCNBC4qcxvWzNU8NGzOnGyK/ZqHhjngmLU8eQab3YC3Cj6/ndKK5omIlcBiYMcS65Yq8xxgQkS8W+Y21iFpuKRqSdXz588vcxctT5deuvbplDWWLcvSzaxlaRUDBCTtCpwE3LChZUTEqIiojIjKioqKhqucbbC5c9cv3cyarzyDzTxgj4LPu6e0onkktQO2AxaWWLeu9IOAvYGZkmYDHSTNrGcb1sz5oWFmrUeewWYysI+krpI2J+vwn1ArzwTgtPR+MPBYZM+pngAMSSPJugL7AM/VVWZE/CkiPh8RXSKiC7AsDQgotQ1r5vzQMLPWI7fn2UTESknnAA8BbYFbI+JlSVcA1RExARgN3JWuQt4nCx6kfOOAV4CVwNkRsQqgWJn1VKXoNqz5q3l+y6WXZk1nnTplgcbPdTFreeST/M+qrKyM6urqpq6GmVmLImlKRFQWW9YqBgiYmVnz5mBjZma5c7AxM7PcOdiYmVnuHGzMzCx3DjZmZpY7BxszM8udg42ZmeXOwcbMzHLnYGNmZrlzsDEzs9w52JiZWe4cbMzMLHcONmZmljsHGzMzy52DjZmZ5c7BxszMcudgY2ZmuXOwMTOz3DnYmJlZ7hxszMwsdw42ZmaWOwcbMzPLXa7BRtJASTMkzZR0cZHlW0i6Jy1/VlKXgmWXpPQZkgbUV6ak0ZKmS3pB0nhJW6f0YZLmS5qWXmfmuc9mZvZZuQUbSW2BG4HjgG7AKZK61cp2BrAoIvYGrgWuTut2A4YA3YGBwE2S2tZT5nkRcWBE9ALmAucUbOeeiOidXr/LY3/NzKxueV7Z9AVmRsSsiPgUGAsMqpVnEHBHej8eOFqSUvrYiFgeEW8CM1N5dZYZER8CpPXbA5Hjvn1WVRV06QJt2mR/q6oadfNmZs1ZnsFmN+Ctgs9vp7SieSJiJbAY2LHEuiXLlHQb8B6wP3BDQb4TC5rX9ihWWUnDJVVLqp4/f37ZOwlkgWX4cJgzByKyv8OHO+CYmSWtaoBARJwO7Aq8Cpyckv8H6JKa1/6XtVdStdcdFRGVEVFZUVGxfhu+9FJYtmzdtGXLsnQzMysv2CjzLUk/T587Sepbz2rzgMKriN1TWtE8ktoB2wELS6xbb5kRsYqsee3E9HlhRCxPi38HHFxPvdff3Lnrl25mtokp98rmJqAfcEr6vISso76UycA+krpK2pysw39CrTwTgNPS+8HAYxERKX1IGq3WFdgHeK6uMlMw3BvW9NmcALyWPu9SsL0TyK56GlanTuuXbma2iWlXZr5DI6KPpOcBImJR+rGvU0SslHQO8BDQFrg1Il6WdAVQHRETgNHAXZJmAu+TBQ9SvnHAK8BK4Ox0xUIdZbYB7pC0LSBgOvD9VJUfSjohlfM+MKzMfS7fiBFZH01hU1qHDlm6mZmh7EKinkzSs8AXgMkp6FQAD0fEQXlXsClUVlZGdXX1+q1UVZX10cydm13RjBgBQ4fmU0Ezs2ZI0pSIqCy2rNwrm+uB+4HPSRpB1uT1swaqX+swdKiDi5lZHcoKNhFRJWkKcDRZM9W/RETD932YmVmrVDLYSNqh4OM/gTGFyyLi/bwqZmZmrUd9VzZTyO7EF9AJWJTedySbEqZrrrUzM7NWoeTQ54joGhF7Ao8AX4+InSJiR+BrwMONUUEzM2v5yr3P5rCImFjzISIeJBudZmZmVq9yR6O9I+lnwN3p81DgnXyqZGZmrU25VzanABVkw5/vBz7H2tkEzMzMSip36PP7wI9yrouZmbVSZQUbSY9T5PkwEfHlBq+RmZm1OuU2o10AXJhe/wFMA9ZzPhczM6tTK38AY7nNaFNqJf1N0nM51MfMbNNT8wDGmsl8ax7ACK1mGqxyn2ezQ8FrJ0kDyJ49Y2ZmG2sTeABjuUOfC2cSWAm8CZyRV6XMzDYpm8ADGMsNNgdExCeFCZK2yKE+Zmabnk6dsqazYumtRLkDBJ4qkvZ0Q1bEzGyTNWJE9sDFQq3sAYz1zfr8eWA3oL2kg8ia0QC2BTrUuaKZmZWvZhBAK34AY33NaAPIHqO8O3BNQfoS4N9zqpOZ2aanlT+AsWSwiYg7gDsknRgRv2+kOpmZWStTXzPatyLibqCLpPNrL4+Ia4qsZmZmto76Bghslf5uDWxT67V1jvUyM9vktaZJBeprRrslvX0kIv5WuEzSF3OrlZnZJq61TSpQ7tDnG8pMMzOzBtDaJhUoGWwk9ZP0E6BC0vkFr8uBtvUVLmmgpBmSZkq6uMjyLSTdk5Y/K6lLwbJLUvqMND1OyTIljZY0XdILksZL2rq+bZiZNVetbVKB+q5sNifrm2nHuv01HwKDS60oqS1wI3Ac0A04RVK3WtnOABZFxN7AtcDVad1uwBCgOzAQuElS23rKPC8iDoyIXsBc4JxS2zAza87qmjygpU4qUDLYRMRfIuIXwGER8YuC1zUR8fd6yu4LzIyIWRHxKTAWGFQrzyDgjvR+PHC0JKX0sRGxPCLeBGam8uosMyI+BEjrt2ft83fq2hhKzDcAABKmSURBVIa1YK2p49SsmNY2qUC5fTbLJI2UNFHSYzWvetbZDXir4PPbKa1onohYCSwGdiyxbskyJd0GvAfsz9o+pbq2sQ5JwyVVS6qeP39+PbtmTamm43TOHIhY23HqgGOtydChMGoUdO4MUvZ31KiWOTgAyg82VcBrQFfgF8BsYHJOddpgEXE6sCvwKnDyeq47KiIqI6KyoqIil/pZw2htHadmdRk6FGbPhtWrs78tNdBA+cFmx4gYDaxITWvfAep7JPQ8YI+Cz7untKJ5JLUje0bOwhLr1ltmRKwia147sZ5tWAvV2jpOzTYF5QabFenvu5K+mibl3KGedSYD+0jqKmlzsg7/CbXyTABOS+8HA49FRKT0IWkkWVdgH+C5uspUZm9Y02dzAtmVWKltWAvV2jpOzTYF5T7P5kpJ2wE/IesL2Rb4cakVImKlpHOAh8iGSd8aES9LugKojogJwGjgLkkzgffJggcp3zjgFbKHtZ2drlioo8w2ZHO4bUs2M/V04PupKkW3YS3XiBHr3uwGLbvj1GxToA09yZf044i4roHr0yxUVlZGdXV1U1fDSqiqatWzsZu1SJKmRERlsWXlNqMV85mJOc0aS+2OU/BQaLPmrNxmtGJ8r4o1C61tDimz1mhjrmzcyW7NgodCmzV/9T3PZgnFg0rNXfpmTc5Doc2av/oeMbBNY1XEbEN16pQ1nRVLN7PmYWOa0cyahdY2h5RZa+RgYy1ea5tDyqw12pjRaGbNxtChDi5mzZmvbMxK8KMMzBqGr2zM6uD7d8wajq9szOrg+3fMGo6DjVkdfP+OWcNxsDGrgx9lYNZwHGzM6uD7d8wajoONWR18/45Zw/FoNLMSfP+OWcPwlY2ZmeXOwcaskfgGUduUuRnNrBH4BlHb1PnKxhqcz+A/yzeI2qbOVzbWoHwGX5xvELVNna9srEH5DL443yBqeWhJrQgONtagfAZfnG8QtYZW04owZw5ErG1FaK4BJ9dgI2mgpBmSZkq6uMjyLSTdk5Y/K6lLwbJLUvoMSQPqK1NSVUp/SdKtkjZL6UdJWixpWnr9PM993tT5DL443yBqDa2ltSLkFmwktQVuBI4DugGnSOpWK9sZwKKI2Bu4Frg6rdsNGAJ0BwYCN0lqW0+ZVcD+QE+gPXBmwXYmRUTv9Lqi4ffWavgMvm5Dh8Ls2bB6dfbXgcY2RktrRcjzyqYvMDMiZkXEp8BYYFCtPIOAO9L78cDRkpTSx0bE8oh4E5iZyquzzIiYGAnwHLB7jvtmdfAZvFnjaGmtCHkGm92Atwo+v53SiuaJiJXAYmDHEuvWW2ZqPvs28OeC5H6Spkt6UFL3YpWVNFxStaTq+fPnl7eHVpTP4M3y19JaEVrjAIGbgCcjYlL6PBXoHBEHAjcADxRbKSJGRURlRFRWVFQ0UlXNzDZMS2tFyPM+m3nAHgWfd09pxfK8LakdsB2wsJ516yxT0mVABfDdmrSI+LDg/URJN0naKSIWbOB+mZk1Cy1potg8r2wmA/tI6ippc7IO/wm18kwATkvvBwOPpT6XCcCQNFqtK7APWT9MnWVKOhMYAJwSEatrNiDp86kfCEl9yfZ5YS57bGZmReUWbFIfzDnAQ8CrwLiIeFnSFZJOSNlGAztKmgmcD1yc1n0ZGAe8Qtb3cnZErKqrzFTWzcDOwNO1hjgPBl6SNB24HhiSApo1gZZ0E5qZNRz5d/ezKisro7q6uqmr0erUnsoGsg7N5tzObGblkzQlIiqLLWuNAwSsmWppN6GZWcNxsLFG09JuQjOzhuNgY42mpd2EZmYNx8HGGk1LuwnNzBqOg401mpZ2E5qZNRw/PM0aVUu6Cc3MGo6vbMzMLHcONmZmljsHGzMzy52DjZmZ5c7BxszMcudgY2ZmuXOwMTOz3DnYmJlZ7hxszMwsdw42ZmaWOwcbMzPLnYONmZnlzsHGzMxy52BjZma5c7AxMzOqqqBLF2jTJvtbVdWw5ft5NmZmm7iqKhg+HJYtyz7PmZN9hoZ7/pSvbJqTvE8tzMyKuPTStYGmxrJlWXpDyTXYSBooaYakmZIuLrJ8C0n3pOXPSupSsOySlD5D0oD6ypRUldJfknSrpM1SuiRdn/K/IKlPnvu8wWpOLebMgYi1pxYOOGaWs7lz1y99Q+QWbCS1BW4EjgO6AadI6lYr2xnAoojYG7gWuDqt2w0YAnQHBgI3SWpbT5lVwP5AT6A9cGZKPw7YJ72GA79t+L1tAI1xamFmVkSnTuuXviHyvLLpC8yMiFkR8SkwFhhUK88g4I70fjxwtCSl9LERsTwi3gRmpvLqLDMiJkYCPAfsXrCNO9OiZ4COknbJa6c3WGOcWpiZFTFiBHTosG5ahw5ZekPJM9jsBrxV8PntlFY0T0SsBBYDO5ZYt94yU/PZt4E/r0c9kDRcUrWk6vnz55exew2sMU4tzMyKGDoURo2Czp1Byv6OGtVwgwOgdQ4QuAl4MiImrc9KETEqIiojorKioiKnqpXQGKcWZmZ1GDoUZs+G1auzvw0ZaCDfYDMP2KPg8+4prWgeSe2A7YCFJdYtWaaky4AK4Pz1rEfTa4xTCzOzJpJnsJkM7COpq6TNyTr8J9TKMwE4Lb0fDDyW+lwmAEPSaLWuZJ37z5UqU9KZwADglIhYXWsbp6ZRaYcBiyPi3Tx2eKPlfWphZtZEcrupMyJWSjoHeAhoC9waES9LugKojogJwGjgLkkzgffJggcp3zjgFWAlcHZErAIoVmba5M3AHODpbIwB90XEFcBE4HiyQQbLgNPz2mczMytO2YWEFaqsrIzq6uqmroaZWYsiaUpEVBZb1hoHCJiZWTPjYGNmZrlzsDEzs9w52JiZWe4cbMzMLHcONmZmljsHGzMzy52DjZmZ5c7BxszMcudgY2ZmuXOwMTOz3DnYmJlZ7hxszMwsdw42ZmaWOwcbM7OGVFUFXbpAmzbZ36qqpq5Rs5Dbw9PMzDY5VVUwfDgsW5Z9njMn+wyb/JN3fWVjZtZQLr10baCpsWxZlr6Jc7AxM2soc+euX/omxMHGzKyhdOq0fumbEAcbM7OGMmIEdOiwblqHDln6Js7BxsysoQwdCqNGQefOIGV/R43a5AcHgEejmZk1rKFDHVyK8JWNmZnlLtdgI2mgpBmSZkq6uMjyLSTdk5Y/K6lLwbJLUvoMSQPqK1PSOSktJO1UkH6UpMWSpqXXz/PbYzMzKya3ZjRJbYEbgWOAt4HJkiZExCsF2c4AFkXE3pKGAFcDJ0vqBgwBugO7Ao9I2jetU1eZfwP+CDxRpDqTIuJrDb6TZmZWljyvbPoCMyNiVkR8CowFBtXKMwi4I70fDxwtSSl9bEQsj4g3gZmpvDrLjIjnI2J2jvtjZmYbKM9gsxvwVsHnt1Na0TwRsRJYDOxYYt1yyiymn6Tpkh6U1H19dsLMzDbepjAabSrQOSKWSjoeeADYp3YmScOBNIkRSyXNaMQ6NoadgAVNXYlmzMenNB+f0nx8Mp3rWpBnsJkH7FHwefeUVizP25LaAdsBC+tZt74y1xERHxa8nyjpJkk7RcSCWvlGAaPq26mWSlJ1RFQ2dT2aKx+f0nx8SvPxqV+ezWiTgX0kdZW0OVmH/4RaeSYAp6X3g4HHIiJS+pA0Wq0r2ZXIc2WWuQ5Jn0/9QEjqS7bPCxtkD83MrCy5XdlExEpJ5wAPAW2BWyPiZUlXANURMQEYDdwlaSbwPlnwIOUbB7wCrATOjohVkA1xrl1mSv8hcBHweeAFSRMj4kyyIPZ9SSuBj4EhKaCZmVkjkX93Nw2ShqemQivCx6c0H5/SfHzq52BjZma583Q1ZmaWOwcbMzPLnYNNC9LIc83dLunNgjnleue9fxurkY+PJI2Q9LqkV9MAlWarkY/NpIJ/N+9IeiDv/dtYjXx8jpY0NR2fv0raO+/9axYiwq8W8CIbffcGsCewOTAd6FYrzw+Am9P7IcA96X23lH8LoGsqp22pMoHbgcFNvd/N+PicDtwJtEmfP9fUx6C5HJta5f4eOLWpj0FzOj7A68ABBeXe3tTHoDFevrJpORp1rrkWqLGPz/eBKyJiNUBE/DPHfdtYTfJvR9K2wJfJZu1ozhr7+ASwbXq/HfBOTvvVrDjYtBxNMdfcCEkvSLpW0hYNsRM5auzjsxfZDOXVyubc+8wUSM1IU81T+C/Ao1Ewi0cz1djH50xgoqS3gW8DVzXIXjRzDjZWl0uA/YFDgB2AnzZtdZqdLYBPIpui5P8BtzZxfZqjU4AxTV2JZug84PiI2B24DbimievTKBxsWo71mWsOlTfXXJ1lRsS7kVlO9h+ib4PtST4a9fiQnanel97fD/Ta6D3IT2MfG5Q9wLAv8KcG2YN8NdrxkVQBHBgRz6b0e4AvNMxuNHNN3WnkV3kvsqmFZpF1QtZ0OHavleds1u3EHJfed2fdTsxZZB2YdZYJ7JL+CrgOuKqpj0EzOz5XAd9J748CJjf1MWguxyat9z3gjqbe9+Z2fFL6AmDftP4ZwO+b+hg0ynFu6gr4tR5fFhxPNpLlDeDSlHYFcEJ6vyVwL1kn5XPAngXrXprWmwEcV6rMlP4Y8CLwEnA3sHVT738zOz4dyc7aXwSeJjtbbfJj0ByOTVr2BDCwqfe7OR4f4Bvp3830dJz2zHv/msPL09WYmVnu3GdjZma5c7AxM7PcOdiYmVnuHGzMzCx3DjZmZo1M0kmSXpa0WlJlHXm2lPScpOkp7y8KlhWdCFbS9pLuTzN/PCepR0rfQ9Ljkl5JZf2o1rbOlfRaWvZ/66l7b0lPp7wvSDq5nH12sDFrYJKW5lz+REkd0+sHG7D+UZL+mEfd7LPS8b69VvJLwDeBJ0usuhz4ckQcCPQGBko6LC0bRnbT6P4RcQDZ3GsA/w5Mi4hewKnAr1P6SuAnEdENOAw4W1K3VL/+ZPO2HRgR3YH/rGeXlpFNrtodGAhcJ6ljPes42Ji1NBFxfER8QHavz3oHG2t6EfFqRMyoJ09ERM2Jy2bpVXOvSl0TwXYju0eOiHgN6CJp58hmBJma0pcAr7J2rrbvk920vbywLEltJY2UNDldwXw3LX89Iv6e3r8D/BOoqG+fHWzMGkFqengm/ae9X9L2Kf0JSVenJo/XJR2e0jtIGpeaPe5Pz1CpTMtmp+lgrgL2Ss9FGVn7ikXSbyQNS+8HpmaSqWRn1DV5tpJ0a9r+85Ja6qzfrVL6wZ9G9oP+v7F2mpu6JoKdTvp+JfUFOpNNlVNYZhfgIKCmrH2Bw9O/sb9IOiSlnwEsjohDyOZIPEtS11pl9SWbIeGN+vbFwcascdwJ/DQ1b7wIXFawrF1E9AV+XJD+A2BRavb4D+DgImVeDLwREb0j4sK6NixpS7LJQr+eyvl8weJLgcfS9vsDIyVttSE7aOtKP97TgN8BJ2jtA+UG1LdujYhYFRG9yQJG35o+GOqeCPYqoGPa7rnA88CqgjptTfaMoR/H2tm425FNtnsYcCEwLj0+4Vjg1FTWs2SzXO9TUNYuwF3A6TVXWKW0K3enzWzDSNoO6BgRf0lJd5BNfVKjZkLPKUCX9P5LpPb2iHhJ0gsbUYX9gTdrmj4k3Q0MT8uOJfshvCB93hLoRNbMYhshIg6FrM8GGBYRwzairA8kPU7WR/ISn50I9raU70OyB/uRAsabZHO0IWkzskBTFRH3FRT/NnBfZNPJPCdpNbAT2byI50bEQ7Xro+xZRX8im4bnmXL2wVc2Zk1vefq7io07AVzJuv+ntyxjHQEnpquj3hHRKSIcaJoBSRU1He+S2gPHAK+lxQ+QXYkCHEk2Bxtp0MjmKf1M4MmI+DAFntHAqxFR+5EGa8qStC9Zs9gC4CHg+ylIIWnf1Oy6OVmAuzMixpe7Pw42ZjmLiMXAopr+GLIHZv2lxCoAfwP+FSCNGupZJM8SYJuCz3OAbpK2SD9SR6f0mo7ivdLnUwrWeQg4N/0YIemg8vbKNoakbyh7eFo/4E+SHkrpu0qamLLtAjyermonk/XZ1PTJXQWcKOlF4FdkgQXgAOAlSTOA44CaIc5fJPt39+WC5rzj07JbgT0lvUQ2qu20dJXzO+AVYGpadgvZydC/AkcAwwrK6l3vPnsiTrOGlZohCh/1ew3ZCKGbgQ5kzRqnR8QiSU8AF0REder0r46ILqnf5A6y0UWvkT3L/qSI+Luk2UBlRCyQ9N9kz9J5MCIuVHaPxDfImk+WAhMi4nZJA8keFbEMmATsFRFfS2fM15E9U6UNWXPb13I8PLaJcrAxa4YktQU2i4hP0hXJI8B+kT3P3qzF8QABs+apA1kTymZk/So/cKCxlsxXNmZmljsPEDAzs9w52JiZWe4cbMzMLHcONmZmljsHGzMzy93/B0Xr3YrPZJdsAAAAAElFTkSuQmCC\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "T6F4VBZn6YGk",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 295
        },
        "outputId": "5fe8ee17-eee7-4556-810f-82d28ee958fe"
      },
      "source": [
        "lat_RP = set(train_semi['Latitude_OP'])\n",
        "long_RP = set(train_semi['Longitude_OP'])\n",
        "lat_AP = set(train_semi['Latitude_AP'])\n",
        "long_AP = set(train_semi['Longitude_AP'])\n",
        "plt.scatter(x=list(long_RP),y=list(lat_RP), color=\"blue\", label=\"Reference Points\")\n",
        "plt.scatter(x=list(long_AP),y=list(lat_AP), color=\"red\", label=\"Access Points\")\n",
        "plt.title(\"Visualising Training Dataset\")\n",
        "plt.xlabel(\"Longitude\")\n",
        "plt.ylabel(\"Latitude\")\n",
        "plt.legend(loc='best',fancybox=True, shadow=True)\n",
        "plt.show()"
      ],
      "execution_count": 27,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "57EH3jrW7TcJ",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 295
        },
        "outputId": "98c22b19-6e60-487e-d121-2d175342f7b0"
      },
      "source": [
        "lat_RP_tr = set(train_semi['Latitude_OP'])\n",
        "long_RP_tr = set(train_semi['Longitude_OP'])\n",
        "lat_AP_tr = set(train_semi['Latitude_AP'])\n",
        "long_AP_tr = set(train_semi['Longitude_AP'])\n",
        "lat_RP_te = set(test_semi['Latitude_OP'])\n",
        "long_RP_te = set(test_semi['Longitude_OP'])\n",
        "lat_AP_te = set(test_semi['Latitude_AP'])\n",
        "long_AP_te = set(test_semi['Longitude_AP'])\n",
        "plt.scatter(x=list(long_RP_tr),y=list(lat_RP_tr), color=\"blue\", label=\"Training Points\")\n",
        "plt.scatter(x=list(long_RP_te),y=list(lat_RP_te), color='xkcd:light green', label=\"Testing Points\")\n",
        "plt.scatter(x=list(long_AP),y=list(lat_AP), color=\"red\", label=\"Access Points\")\n",
        "plt.title(\"Visualising the Dataset\")\n",
        "plt.xlabel(\"Longitude\")\n",
        "plt.ylabel(\"Latitude\")\n",
        "plt.legend(loc='best',fancybox=True, shadow=True)\n",
        "plt.show()"
      ],
      "execution_count": 28,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAZsAAAEWCAYAAACwtjr+AAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nO3de3hV1Z3/8fcnCSCIglWqgmJQAeUaJYKXUXAsQtVKragwsUqrotZL61S8lNaqvzKjo492xFKl3m0KWLxRxToyVsGqQNBgxQJGDQhah4sGECskfH9/7JV4Ek+Sk5Cdc3L4vp4nT85Ze++110og37Muey2ZGc4551ycctJdAOecc9nPg41zzrnYebBxzjkXOw82zjnnYufBxjnnXOw82DjnnIudBxvXpklaJmlEzPcwSYeG1/dI+kUmlCvhXg9J+lVr3Mu55vJg4zKWpD9LujlJ+hhJ/5CUZ2b9zeyl1iqTmV1iZv8vhfNiKZekCZJeael8E/J/SdI/JW2WtEnSEknXSerQhDxqgnOcWus+rmV4sHGZ7GHgXEmqk/59oNjMKtNQpl3B5Wa2B7A/8FNgHDA3ye/BuZR5sHGZ7Clgb+D46gRJewGnAY+E9+WSvhVeD5VUEj6RfyLpjpA+QtKaxIyTXPeapM8kfSzpbkntkxUosctK0j6SngnXbZS0QFJOkvxvlPSYpEdCi2GZpMKEPI+U9GY49kdJs5J1i0k6HLgHOEbSFkmfJRzeS9KzIY+Fkg5JuO4wSS+EMq6QdHYqP3wz+zy0zk4HjgFObeznJWl+uHxpKOM5kvYKP6d1kj4Nrw9IKN8ESe+Hsn8gqSjh2A8l/T1c97ykg+q7Typ1cunjwcZlLDP7AngMOC8h+WxguZktTXLJfwP/bWZ7AoeEa1NRBVwF7EP0R/Uk4EcpXPdTYA3QDdgX+BlQ3/pPpwMzga7AHOBugPBH+kngIeAbwAzgjGQZmNnfgUuA18yss5l1TTg8DrgJ2AsoA6aE/HcHXgD+AHwznDdNUr8U6ld939VACV8F/Xp/XmZ2QjhncCjjLKK/Mw8CBwE9gS8S6r87cBfw7dCaOhYoDcfGEP1Mv0f0M14Qfj713cdlMA82zSDp/0l6S1KppP+R1L2e86rCOaWS5iSkXy6pLPQ575OQPkJSRcI1N4T0vglppeGT+08aKeMJkt6QVClpbEvVPQ0eBsZK2i28Py+kJbMdOFTSPma2xcxeT+UGZrbEzF43s0ozKwfuBYancOl2oq6mg8xsu5ktsPoXG3zFzOaaWRXwKDA4pB8N5AF3hTyeABalUu46njSzRaFrsRgoCOmnAeVm9mCo35vA48BZTcz/I6Jg2OSfl5ltMLPHzWyrmW0mCoSJ5+8ABkjqaGYfm9mykH4J8J9m9vdQr/8ACqpbN65t8WDTiBAAHqqTfJuZDTKzAuAZ4IZ6Lv/CzArC1+kJ6X8FvgWsSnLNgoRrbgYwsxXVacAQYCvRp+GGrAYmEH2ibbPM7BVgPfDd0DU0lPrrdAHQB1guabGk01K5h6Q+oWvnH5I2Ef1R26ex64DbiFoR/xO6ga5r4Nx/JLzeCuwmKQ/oDqytE6Q+TKXcjeTfObw+CBgWurw+C11vRcB+Tcy/B7ARmv7zktRJ0r2SVoXz5wNdJeWa2efAOUSB5ePQFXhYQtn/O6HcGwGFsrg2xoNNM5jZpoS3u1N/10l9178ZPhE2x0nAe2a2CkDSIYpmbS0JYwaHhXuUm9lbRJ8a27pHiFo05wLPm9knyU4ys3fNbDxRd9GtwOzQTfM50Kn6PEm5RN0y1X4LLAd6hy64nxH9UWuQmW02s5+a2cFE3WT/LumkJtbtY6CHVGvw/cCGbtvE/D8EXjazrglfnc3s0lQzkHQg0YecBSGpqT+vnwJ9gWHh/OouMAGY2fNmNpKolbgc+F1C2S+uU/aOZvZqqmV3mcODTTNJmiLpQ6JPifW1bHZTNGD9uqTvppj1MZKWSnpOUv8kx8cR+q2D6cAVZjYEuBqYlmod2pBHiFqCF1F/FxqSzpXUzcx2ANWD5zuAlUS/i1MltQN+DiRO5d0D2ARsCcE6pT/Ekk6TdGgIFBVEYxlNDe6vhesul5QXximGNnD+J8ABqmcCQxLPAH0kfV9Su/B1lKLJBg0KLZLhwNNEXXtzw6HGfl6fAAcnvN+DaJzmM0nfAH6ZcI99FU1l3x34EtjCVz/De4Drq/8fSOoiKbH7r+59XAbzYFMPRTN6SoH7gNMTxktGAZjZZDM7kKh//PJ6sjnIzAqBfwN+rYQZQvV4I1wzGJhKNBsrsUztiT5B/zG870w0oPrHUNZ7iT4dZpXQCnyVqBU5p4FTRwPLJG0hmiwwzsy+MLMKogHs+4C1RC2dxNlpVxP9jjYTfapOdbC5NzCP6A/ka8A0M/tLitcCYGbbiAbALyAKkOcSBYgv67nkRWAZ8A9J61PIfzNwMtGHlI+IuttupXawretuSZuJ/pj/mmiMZ3QI4tD4z+tG4OHQ/XV2yKMjUXfo68CfE87NAf49lG0j0VjOpaHsT4ayzgzdb28D327gPi6DyTdPa5iip8AnmNmEeo73BOaa2YBG8nkIeMbMZieklQOFZpb0j0bd4+FT72VmdnJ4vyewwszqDTDJ7usym6SFwD1m9mC6y+JcS/GWTTNI6p3wdgxRP3Pdc/ZSeOo6zDg7DninkXz3q+67lzSU6PezIeGU8SR0oYWxow+quxYUGYxrUyQND7/7PEnnA4Oo/enfuTbPg03z3CLpbUlvEXVR/BhAUqGk+8I5hwMlkpYCfwFuMbN3wnlXKnrI8ADgrYRrxgJvh2vuIuoGsnDN7sBI4Ik6ZSkCLgjXLCMKfoR++TVEU1zvlbQMl6n6AkuJutF+Cow1s4/TWyTnWpZ3oznnnIudt2ycc87FLi/dBchE++yzj+Xn56e7GM4516YsWbJkvZl1S3bMg00S+fn5lJSUpLsYzjnXpkhKtioK4N1ozjnnWoEHG+ecc7HzYOOccy52PmbjnEu7bdu28d5777F169Z0F8WloFOnThxyyCG0b5/qEn0ebJxzGeC9996ja9eu9O3bl5wc73DJZDt27OCTTz6hrKyMfv1S3oPPg43LbCurSllY9QJbqKAzXRiWO5I+uQWNX+jalK1bt3qgaSNycnLYd999Wbt2La+++irHHHMMtXfIqOe6Viibc82ysqqUl6ueZgsVAGyhgpernmZlVWmaS+bi4IGm7cjJyUESr7/+OqtW1TvbufY1MZfJuWZbWPUClWyvlVbJdhZWvZCmEjnnEkli8+bNKZ3rwcZlrOoWTarpzjXXhg0bKCgooKCggP32248ePXrUvN+2bVuD15aUlHDllVc2eo9jjz22Rcr60ksv0aVLFwoKCjj88MO56aabGjz/hhtuYN68eY3m+eqr8W6A6mM2LmN1pkvSwNKZLmkojctme++9N6WlUffsjTfeSOfOnbn66qtrjldWVpKXl/zPZWFhIYWFhY3eoyX/mB9//PE888wzfP755xQUFPCd73yHI488Mum5N998c6P5vfTSS3Tu3LnFAmIy3rJxGWtY7kjyaFcrLY92DMsdmaYSuUxRXAz5+ZCTE30vLm75e0yYMIFLLrmEYcOGcc0117Bo0SKOOeYYjjjiCI499lhWrFgBRH+oTzvtNCAKVD/84Q8ZMWIEBx98MHfddVdNfp07d645f8SIEYwdO5bDDjuMoqIiqlffnzt3LocddhhDhgzhyiuvrMm3PrvvvjtDhgyhrKyM0tJSjj76aAYNGsQZZ5zBp59+WlOP2bOjvRPz8/P55S9/yZFHHsnAgQNZvnw55eXl3HPPPdx5550UFBSwYMEC/vjHPzJgwAAGDx7MCSec0CI/T2/ZuIxVPevMZ6O5RMXFMHEiVD+Ss2pV9B6gqKhl77VmzRpeffVVcnNz2bRpEwsWLCAvL4958+bxs5/9jMcff/xr1yxfvpy//OUvbN68mb59+3LppZfSrl3tD01vvvkmy5Yto3v37hx33HH89a9/pbCwkIsvvpj58+fTq1cvxo8f32j5NmzYwOuvv84vfvELxo8fz9SpUxk+fDg33HADN910E7/+9a+/ds0+++zDG2+8wbRp07j99tu57777uOSSS2q15gYOHMjzzz9Pjx49+Oyzz5r506vNg43LaH1yCzy4uFomT/4q0FTbujVKb+lgc9ZZZ5GbmwtARUUF559/Pu+++y6S2L59e9JrTj31VDp06ECHDh345je/ySeffMIBBxxQ65yhQ4fWpBUUFFBeXk7nzp05+OCD6dWrFwDjx49n+vTpSe+xYMECjjjiCHJycrjuuus44IAD+Oyzzxg+fDgA559/PmeddVbSa7/3ve8BMGTIEJ54ou5ejJHjjjuOCRMmcPbZZ9ecv7M82Djn2pTVq5uWvjN23333mte/+MUvOPHEE3nyyScpLy9nxIgRSa/p0KFDzevc3FwqKyubdU5DqsdsqlVUpD5ppvreDd33nnvuYeHChTz77LMMGTKEJUuWsPfeezepjHV5sHFZxR8CzX49e0ZdZ8nS41RRUUGPHj0AeOihh1o8/759+/L+++9TXl5Ofn4+s2bNSvnaLl26sNdee7FgwQKOP/54Hn300ZpWTir22GMPNm3aVPP+vffeY9iwYQwbNoznnnuODz/8cKeDjU8QcFljV3wItDUGyjPNlCnQqVPttE6dovQ4XXPNNVx//fUcccQRTW6JpKJjx45MmzaN0aNHM2TIEPbYYw+6dEl95uXDDz/MpEmTGDRoEKWlpdxwww0pX/ud73yHJ598smaCwKRJkxg4cCADBgzg2GOPZfDgwc2pUi2qngXhvlJYWGi+eVrb8+i22+qdKv399pPSUKJ41R0oh+iP7vTpLT92EbclS5YwZMiQlM8vLo7GaFavjlo0U6a0vTons2XLFjp37oyZcdlll9G7d2+uuuqqdBcrqSVLljB//ny+9a1vMXDgQAAkLTGzpPPAvWXjssau9hBoQwPl2a6oCMrLYceO6Hs2BBqA3/3udxQUFNC/f38qKiq4+OKL012kFuNjNi5r7GoPgbbmQLlrHVdddVXGtmR2VqwtG0mjJa2QVCbpuiTHO0iaFY4vlJSfcOz6kL5C0qgm5HmXpC0J7ydIWiepNHxd2PI1dZlgV3sItL4B8bgHyp1rjtiCjaRc4DfAt4F+wHhJdTc/uAD41MwOBe4Ebg3X9gPGAf2B0cA0SbmN5SmpENgrSXFmmVlB+LqvJevpMkef3AKG546pacl0pgvDc8dk7Wy0dA2UO9cccXajDQXKzOx9AEkzgTHAOwnnjAFuDK9nA3cr2hhhDDDTzL4EPpBUFvKjvjxDILoN+DfgjBjr5TLYrvQQaPU4RTYOlLvsE2c3Wg/gw4T3a0Ja0nPMrBKoAPZu4NqG8rwcmGNmHycpy5mS3pI0W9KByQoraaKkEkkl69atS6V+zqVdtg6Uu+yTFbPRJHUHzgKmJjn8JyDfzAYBLwAPJ8vDzKabWaGZFXbr1i2+wjrnMs7ObDEAX1+i/5577uGRRx5pkbKNGDGCvn37MnjwYI477riaBUCT+eijjxg7dmyjef7Hf/xHi5StKeIMNmuBxFbEASEt6TmS8oAuwIYGrq0v/QjgUKBMUjnQKXS9YWYbQnccwH1A6pP5nXO7hOotBkpLS7nkkku46qqrat63b9++0evrBptLLrmE8847r8XKV1xczNKlSzn//POZNKn+Z8a6d+9es8JzQ7It2CwGekvqJak90YD/nDrnzAHOD6/HAi9a9JTpHGBcmK3WC+gNLKovTzN71sz2M7N8M8sHtoZJB0jaP+F+pwN/j6W2zrlWs7KqlEe33cZvt/2cR7fdFssqEUuWLGH48OEMGTKEUaNG8fHHUQ/9XXfdRb9+/Rg0aBDjxo1LukT/jTfeyO233w5ELZNrr72WoUOH0qdPHxYsWADA1q1bOfvss+nXrx9nnHEGw4YNo7GHyU844QTKysowMyZNmsSAAQMYOHBgzdI25eXlDBgwAIiW1Pne977H6NGj6d27N9dccw0A1113HV988QUFBQUUFRXx+eefc+qppzJ48GAGDBjQpGVymiK2CQJmVinpcuB5IBd4wMyWSboZKDGzOcD9wKOhFbKRKHgQznuMaDJBJXCZmVUBJMuzkaJcKen0kM9GYEILV9U514qqlyWq3jK8elkioMUmh5gZV1xxBU8//TTdunVj1qxZTJ48mQceeIBbbrmFDz74gA4dOvDZZ5/RtWvXry3R/7//+7+18qusrGTRokXMnTuXm266iXnz5jFt2jT22msv3nnnHd5++20KChov+5/+9CcGDhzIE088QWlpKUuXLmX9+vUcddRRSfedKS0t5c0336RDhw707duXK664gltuuYW77767ZrO4xx9/nO7du/Pss88CTVvUsylifajTzOYCc+uk3ZDw+p9EYy3Jrp0CfG0SZ7I8k5zTOeH19cD1TSq4cy5jLax6oSbQVKtkOwurXmixYPPll1/y9ttvM3Jk9IxWVVUV++8fdZIMGjSIoqIivvvd7/Ld7343pfwSl/UvLy8H4JVXXuHHP/4xAAMGDGDQoEH1Xl9UVETHjh3Jz89n6tSp3HHHHYwfP57c3Fz23Xdfhg8fzuLFi7+Wx0knnVSzvlq/fv1YtWoVBx5Ye47UwIED+elPf8q1117LaaedxvHHH59SnZrKVxBwzrUprbEskZnRv39/Xnvtta8de/bZZ5k/fz5/+tOfmDJlCn/7298azS+VZf0bUlxcnNLW0/Xdt6F79+nThzfeeIO5c+fy85//nJNOOqlJi3imKitmoznndh31LT/UkssSdejQgXXr1tUEm+3bt7Ns2TJ27NjBhx9+yIknnsitt95KRUUFW7ZsYY899mDz5s1Nusdxxx3HY489BsA777yTUtCqdvzxxzNr1iyqqqpYt24d8+fPZ+jQoY1fGLRr165m87ePPvqITp06ce655zJp0iTeeOONJtUjVd6ycc61KcNyR9Yas4GWX5YoJyeH2bNnc+WVV1JRUUFlZSU/+clP6NOnD+eeey4VFRWYGVdeeSVdu3blO9/5DmPHjuXpp59m6tRkT2B83Y9+9CPOP/98+vXrx2GHHUb//v1T3lLgjDPO4LXXXmPw4MFI4r/+67/Yb7/9arroGjNx4kQGDRrEkUceyXnnncekSZPIycmhXbt2/Pa3v00pj6byLQaS8C0GnGtdTd1iIBs2yauqqmL79u3stttuvPfee3zrW99ixYoVKU21zgRN3WLAWzbOuTYnG5Yl2rp1KyeeeCLbt2/HzJg2bVqbCTTN4cHGOefSYI899mj0uZps4sHGOdds2dCd5VqHBxvnXLO0xsOVLnv41GfnXLM09HClc3V5sHHONUtrPFzpsocHG+dcs7TGw5Wt7amnnkISy5cvT3dRALjxxhtrtjsYMGAAc+bUXcu4tlNOOYXPPvuswXMeeughPvroo5YsZko82DjnmmVY7kjyaFcrraUfrmxtM2bM4F/+5V+YMWNGuotSo3q7gz/+8Y/88Ic/ZMeOHfWeO3fuXLp27dpgfh5snHNtSp/cAobnjqlpyXSmC8Nzx7TO5IDiYsjPh5yc6Htx8U5nuWXLFl555RXuv/9+Zs6cWZNeVVXF1VdfXbNYZvUKAYsXL+bYY49l8ODBDB06lM2bN1NVVcWkSZM46qijGDRoEPfeey8AH3/8MSeccEJNC2XBggVUVVUxYcKEmm0C7rzzzgbLd/jhh5OXl8f69euZMWMGAwcOZMCAAVx77bU15+Tn57N+/XrKy8s5/PDDueiii+jfvz8nn3wyX3zxBbNnz6akpISioiIKCgr44osvuO6662q2TKhetToOPhvNOddsaXm4srgYJk6ErVuj96tWRe9hp/bFfvrppxk9ejR9+vRh7733rlnVYPr06ZSXl1NaWkpeXh4bN25k27ZtnHPOOcyaNYujjjqKTZs20bFjR+6//366dOnC4sWL+fLLLznuuOM4+eSTeeKJJxg1ahSTJ0+mqqqKrVu3Ulpaytq1a3n77bcBGu3+WrhwITk5OWzfvp1rr72WJUuWsNdee3HyySfz1FNPfW0F6nfffZcZM2bwu9/9jrPPPpvHH3+cc889l7vvvpvbb7+dwsJCNmzYwJNPPsny5cuR1GgZdoa3bJxzbcvkyV8Fmmpbt0bpO2HGjBmMGzcOgHHjxtV0pc2bN4+LL76YvLzos/k3vvENVqxYwf77789RRx0FwJ577kleXh7/8z//wyOPPEJBQQHDhg1jw4YNvPvuuxx11FE8+OCD3Hjjjfztb39jjz324OCDD+b999/niiuu4M9//jN77rln0nJVb8p29dVXM2vWLEpKShgxYgTdunUjLy+PoqIi5s+f/7XrevXqVbNHTuLWBom6dOnCbrvtxgUXXMATTzxBp06ddupn2BAPNs65tmX16qalp2Djxo28+OKLXHjhheTn53Pbbbfx2GOP0dS1I82MqVOn1mwp/cEHH3DyySdzwgknMH/+fHr06MGECRN45JFH2GuvvVi6dCkjRozgnnvu4cILL0yaZ/WYzYIFC5q010wq2wvk5eWxaNEixo4dyzPPPMPo0aObVN+m8GDjnGtbevZsWnoKZs+ezfe//31WrVpFeXk5H374Ib169WLBggWMHDmSe++9t+aP9caNG+nbty8ff/wxixcvBmDz5s1UVlYyatQofvvb39Ys379y5Uo+//xzVq1axb777stFF13EhRdeyBtvvMH69evZsWMHZ555Jr/61a9SXtp/6NChvPzyy6xfv56qqipmzJjB8OHDU65r4nYIW7ZsoaKiglNOOYU777yTpUuXNuXH1iQ+ZuOca1umTKk9ZgPQqVOU3kwzZsyoNdAOcOaZZzJjxgymTp3KypUrGTRoEO3ateOiiy7i8ssvZ9asWVxxxRV88cUXdOzYkXnz5nHhhRdSXl7OkUceiZnRrVs3nnrqKV566SVuu+022rVrR+fOnXnkkUdYu3YtP/jBD2pml/3nf/5nSmXdf//9ueWWWzjxxBMxM0499VTGjBmTcl0nTJjAJZdcQseOHXnuuecYM2YM//znPzEz7rjjjtR/aE3kWwwk4VsMONe6mrrFAMXF0RjN6tVRi2bKlJ2aHOCazrcYcM5lv6IiDy5tjI/ZOOeci12swUbSaEkrJJVJui7J8Q6SZoXjCyXlJxy7PqSvkDSqCXneJWlLKvdwzmWOhp6Md5mlOb+r2LrRJOUCvwFGAmuAxZLmmNk7CaddAHxqZodKGgfcCpwjqR8wDugPdAfmSeoTrqk3T0mFwF51ipL0HjFU2bkm8b1gvtKpUyc++eQT9t13X3JyvMMlk+3YsYN//OMfNTPuUhXnmM1QoMzM3geQNBMYAyQGmzHAjeH1bOBuSQrpM83sS+ADSWUhP+rLMwS324B/A85o7B7mMyNcGvleMLUdcsghrFy5krVr1xL9CXCZbPv27axevRpJKX84iDPY9AA+THi/BhhW3zlmVimpAtg7pL9e59oe4XV9eV4OzDGzj+v8Y63vHusTT5I0EZgI0HMn5us7l4qG9oLZFYNN+/bt6dOnD7///e/ZunUru+++e7qL5Bqxbds2cnJy+OY3v5nS+VkxG01Sd+AsYERz8zCz6cB0iKY+t0zJnEvO94L5uvbt23PmmWfy0ksv8dlnnzX56X3Xurp27crRRx9Nt27dUjo/zmCzFjgw4f0BIS3ZOWsk5QFdgA2NXJss/QjgUKAstGo6SSozs0MbuIdzadOZLkkDS1veC6YldOnSpUkPKLq2I86RuMVAb0m9JLUnGvCvu/PPHOD88Hos8GIYS5kDjAszyXoBvYFF9eVpZs+a2X5mlm9m+cDWEGgauodzaZONe8E415DYWjZhfORy4HkgF3jAzJZJuhkoMbM5wP3Ao2ECwEai4EE47zGiyQSVwGVmVgWQLM9GipL0Hs6lU/W4jM9Gc7sKX64mCV+uxjnnmq6h5Wp8QrtzzrnYebBxzjkXOw82zrmsUVwM+fmQkxN9Ly5Od4lctax4zsY554qLa29zs2pV9B58gehM4C0b51xWmDy59n5qEL2fPDk95XG1ebBxzmWF1aublu5alwebFuJ9xc6lV31LGvpSh5nBg00LqO4rXrUKzL7qK/aA41zrmTIFOnWqndapU5Tu0s+DTQvwvmLn0q+oCKZPh4MOAin6Pn26Tw7IFL6CQBJNXUEgJydq0dQlgW8+6JzbVfgKAjHzvmLnnGuYB5sW4H3FzjnXMA82LcD7ip1zrmEebFpIURGUl0djNOXlHmjcrskfAXD18eVqnHMtwpeLcQ3xlo1zrkX4IwCuIR5snHMtwpeLcQ3xYOOcaxH+CIBriAcb51yL8EcAXEM82DjnWoQ/AuAaEmuwkTRa0gpJZZKuS3K8g6RZ4fhCSfkJx64P6SskjWosT0n3S1oq6S1JsyV1DukTJK2TVBq+Loyzzs7tyvwRAFef2IKNpFzgN8C3gX7AeEn96px2AfCpmR0K3AncGq7tB4wD+gOjgWmSchvJ8yozG2xmg4DVwOUJ95llZgXh67446uucc65+cbZshgJlZva+mW0DZgJj6pwzBng4vJ4NnCRJIX2mmX1pZh8AZSG/evM0s00A4fqOgK8wmun8CUDndhlxBpsewIcJ79eEtKTnmFklUAHs3cC1DeYp6UHgH8BhwNSE885M6F47MFlhJU2UVCKpZN26dSlX0jWTbwLk3C4lqyYImNkPgO7A34FzQvKfgPzQvfYCX7Wk6l473cwKzaywW7durVLeXZo/AejcLiXOYLMWSGxFHBDSkp4jKQ/oAmxo4NpG8zSzKqLutTPD+w1m9mU4fB8wpNk1ci3HnwB0bpcSZ7BZDPSW1EtSe6IB/zl1zpkDnB9ejwVetGg3tznAuDBbrRfQG1hUX56KHAo1YzanA8vD+/0T7nc6UavHpZs/AehcrDJtSDS2hTjNrFLS5cDzQC7wgJktk3QzUGJmc4D7gUcllQEbiYIH4bzHgHeASuCy0GKhnjxzgIcl7QkIWApcGopypaTTQz4bgQlx1dk1wZQptVdtBH8C0LkWkomLovq20Ek0dVto10zFxdEYzcaARAMAABlkSURBVOrVUYtmyhR/MMO5FpCfHwWYug46KHr+KS6+LbTLTP4EoMsymdJ1lYlDoh5snHOuBWTSbP5MHBJNKdiEAfhzJd0Q3veUNDTeojnnXNuRSbP5M3FR1FRbNtOAY4Dx4f1momVjnHMuY7Vmt1YmdV1l4qKoqc5GG2ZmR0p6E8DMPg1Tj51zLiO19oysnj2TD8qnq+uqqCizhkFTbdlsD4tgGoCkbsCO2ErlnHM7qbW7tTKx6yqTpBps7gKeBL4paQrwCvAfsZXKOed2Umt3azWl6ypTZq21ppSfs5F0GHAS0UOT/2tmWfskvj9n41zbl65nTRpTt3sPohZQusdUWkKzn7OR9I3qL+D/gBnAH4BPQppzzqVVfa2ETO3WyqRZa62psW60JUBJ+L4OWAm8G14vibdozjnXsIaebcnEGVmwE917bbzvLaVuNEm/A540s7nh/beB75rZxTGXLy28G825tiFTu8oa0qwyt5G+t5ZYrubo6kADYGbPAce2ROGcc665MunZllQ1q3svC/reUg02H0n6uaT88DUZ+CjOgjnnXH2qe5Tq65jJ5J0qmtW91xajah2pPtQ5Hvgl0fRngPl8tZqAc861mmQ9SokyYRJAY5r8wGWmPTHaDCm1bMxso5n92MyOCF8/NrONcRfOOefqStajVC1TJgG0uEydWtcEKbVsJP2FsHpAIjP71xYvkXPONaC+niMpcycF7LTq6NmG939KtRvt6oTXuwFnEu186ZxzrSoLepSaJ9MWO2uilIKNmdV9puavkhbFUB7nnGuQ7yjeNqXajZa4WkAOMAToEkuJnHOuAVnQo7TTVlaVsrDqBbZQQWe6MCx3JH1yC9JdrAal2o22hGjMRkTdZx8AF8RVKOeca0gb71HaKSurSnm56mkq2Q7AFip4ueppgIwOOKkGm8PN7J+JCZI6xFAe55xzDVhY9UJNoKlWyXYWVr2Q0cEm1Yc6X02S9lpjF0kaLWmFpDJJ1yU53kHSrHB8oaT8hGPXh/QVkkY1lqek+yUtlfSWpNmSOjd2D+eca2u2UNGk9EzRYMtG0n5AD6CjpCOIutEA9gQ61XthdG0u0dbRI4E1wGJJc8zsnYTTLgA+NbNDJY0DbgXOkdQPGAf0B7oD8yT1CdfUl+dVZrYp3PsO4HLglvru0dgPxrnW1hb74V3r60yXpIGlc4YPozfWjTYKmAAcANyRkL4Z+Fkj1w4FyszsfQBJM4ExQGKwGQPcGF7PBu6WpJA+08y+BD6QVBbyo748EwKNgI589VxQ0ntYqhv5ONcK2mo/vGt9w3JH1vq3ApBHO4bljkxjqRrXYLAxs4eBhyWdaWaPNzHvHsCHCe/XAMPqO8fMKiVVAHuH9NfrXNsjvK43T0kPAqcQBbSfNnKP9YkFkTQRmAjQM+sn7LtM01b74V3rq/730NZawY11o51rZr8H8iX9e93jZnZHksvSxsx+ELrvphJ1lT3YhGunA9Mh2mIgnhI6l1xb7Yd36dEntyDjg0tdjU0Q2D187wzsUeercyPXrgUOTHh/QEhLeo6kPKJndzY0cG2jeZpZFTCTaJWDhu7hXMaor7890/vhnUtVY91o94aX88zsr4nHJB3XSN6Lgd6SehH9wR8H/Fudc+YA5xPNbBsLvGhmJmkO8Icw0N8d6A0sIpqg8LU8wzjNIWZWFl6fDixv6B6NlN25VtVW++GdS1Wqz9lMBY5MIa1GGB+5HHgeyAUeMLNlkm4GSsxsDnA/8GiYALCRKHgQznuMaOylErgstFioJ88corGlPYkC0lLg0lCUpPdwLpO01X5451LV4LbQko4h2pHzJ8CdCYf2BM4ws8HxFi89fFtol24+Ddq1RQ1tC91Yy6Y90dhMHtE4TbVNRF1SzrkW5tOgXTZqbMzmZeBlSQ+ZWZJFvZ1zLc2nQbtslOqYzVZJtxE90b9bdaJvnuZcy/Np0C4bpbo2WjHR7K5ewE1AOdFsM+dcC/Np0C4bpRps9jaz+4HtZvaymf0Q8FaNczEYljuSPNrVSvNp0K6tS7UbrboD+WNJpwIfAd9o4HznXDP5NGiXjVINNr+S1IVovbGpRFOffxJbqZzbxbXEciQ+fdplkpSCjZk9E15WACcCSPJg41yG8unTLtOkOmaTzNcW5nTOZYaGpk87lw47E2zU+CnOuXTw6dMu0+xMsPHFLJ3LUD592mWaBoONpM2SNiX52ky0GrNzLgP59GmXaRpbrmaPho475zKTT592mSbVqc/OuTamLe7m2FTFxTB5MqxeDT17wpQpUFSU7lK5ZDzYOOdqaSvP5xQXw8SJsHVr9H7Vqug9eMDJRDszQcA5l2Wqn8+pnrVW/XzOyqrSNJfs6yZP/irQVNu6NUp3mcdbNs7tpLbSEkhFW9reYPXqpqW79PKWjXM7oS21BFLRlp7P6dmzaekuvbxl49xOaEstgVR0pkvSwJKJz+dMmVJ7zAagU6cofVeWqS1tb9k4txPaUksgFW3p+ZyiIpg+HQ46CKTo+/Tpu/bkgExuaccabCSNlrRCUpmk65Ic7yBpVji+UFJ+wrHrQ/oKSaMay1NScUh/W9IDktqF9BGSKiSVhq8b4qyz27Vk25P6fXILGJ47pqb8nenC8NwxGfHJOJmiIigvhx07ou+7cqCBzF4TL7ZuNEm5wG+AkcAaYLGkOWb2TsJpFwCfmtmhksYBtwLnSOoHjCPahro7ME9Sn3BNfXkWA+eGc/4AXAj8NrxfYGanxVVXt+saljuy1urKkLktgVTtCs/nZKtMbmnH2bIZCpSZ2ftmtg2YCYypc84Y4OHwejZwkiSF9Jlm9qWZfQCUhfzqzdPM5loALAIOiLFuzgFtryXgslsmt7TjnCDQA/gw4f0aYFh955hZpaQKYO+Q/nqda3uE1w3mGbrPvg/8OCH5GElLiXYYvdrMljWnQs4l4y0BlykyuaWdjbPRpgHzzWxBeP8GcJCZbZF0CvAU0LvuRZImAhMBevrcSedcG5TJa+LFGWzWAgcmvD8gpCU7Z42kPKALsKGRa+vNU9IvgW7AxdVpZrYp4fVcSdMk7WNm6xMLYmbTgekAhYWFvn2Cc65NytSWdpzBZjHQW1IvooAwDvi3OufMAc4HXgPGAi+amUmaA/xB0h1EEwR6E43DqL48JV0IjAJOMrMd1TeQtB/wSch3KNE41YaY6uwckLnPOjiXLrEFmzAGcznwPJALPGBmyyTdDJSY2RzgfuBRSWXARqLgQTjvMeAdoBK4zMyqAJLlGW55D7AKeC2aY8ATZnYzURC7VFIl8AUwLkwicC4W1c86VPebVz/rAHjAcbss+d/dryssLLSSkpJ0F8O1UY9uu63ep/C/335SGkrkXOuQtMTMCpMd8xUEXNYpLob8fMjJib4XF7fu/TP5WQfn0iUbZ6O5XVgm7HHSltYXc661eMvGZZVM2OOkLa0v5lxr8ZaNyyqZsMdJJj/r4Fy6eLBxWaVnz6jrLFl6a8rUZx2cSxfvRnNZZcqUaE+TRL7HiXPp58HGZRXf48S5zOTdaC7rFBV5cHEu03jLxjnnXOw82DjnnIudBxvnnHOx82DjnHMudh5snHPOxc6DjXPOudh5sHHOORc7DzbOOedi58HGOedc7DzYOOeci50vV+Occ46VVaWxbovhwcY553ZxK6tKebnqaSrZDkRbmL9c9TRAiwUc70ZzbUdxMeTnQ05O9L24ON0lci4rLKx6oSbQVKtkOwurXmixe8QabCSNlrRCUpmk65Ic7yBpVji+UFJ+wrHrQ/oKSaMay1NScUh/W9IDktqFdEm6K5z/lqQj46yzi0lxMUycGO2MZhZ9nzjRA45zLWALFU1Kb47Ygo2kXOA3wLeBfsB4Sf3qnHYB8KmZHQrcCdwaru0HjAP6A6OBaZJyG8mzGDgMGAh0BC4M6d8GeoevicBvW762LnaTJ8PWrbXTtm6N0p1zO6UzXZqU3hxxtmyGAmVm9r6ZbQNmAmPqnDMGeDi8ng2cJEkhfaaZfWlmHwBlIb968zSzuRYAi4ADEu7xSDj0OtBV0v5xVdrFZPXqpqU751I2LHckebSrlZZHO4bljmyxe8QZbHoAHya8XxPSkp5jZpVABbB3A9c2mmfoPvs+8OcmlANJEyWVSCpZt25dCtVzrapnz6alO+dS1ie3gOG5Y2paMp3pwvDcMT4brRHTgPlmtqApF5nZdGA6QGFhocVRMLcTpkyJxmgSu9I6dYrSnXM7rU9uQYsGl7ribNmsBQ5MeH9ASEt6jqQ8oAuwoYFrG8xT0i+BbsC/N7EcLtMVFcH06XDQQSBF36dP9/2fnWsj4gw2i4HeknpJak804D+nzjlzgPPD67HAi2HMZQ4wLsxW60U0uL+ooTwlXQiMAsab2Y469zgvzEo7Gqgws4/jqLCLWVERlJfDjh3Rdw80zrUZsXWjmVmlpMuB54Fc4AEzWybpZqDEzOYA9wOPSioDNhIFD8J5jwHvAJXAZWZWBZAsz3DLe4BVwGvRHAOeMLObgbnAKUSTDLYCP4irzs4555JT1JBwiQoLC62kpCTdxXDOuTZF0hIzK0x2zFcQcM45FzsPNs4552LnwcY551zsPNg455yLnQcb55xzsfNg45xzLnYebJxzzsXOg41zzrnYebBxzjkXOw82zjnnYufBxjnnXOw82DjnnIudBxvnnEuX4mLIz4ecnOh7cXG6SxSbbNyp0znnMl9xce3dZ1etit5DVu7V5C0b55xLh8mTa29zDtH7yZPTU56YebBxzrl0WL26aeltnAcb55xLh549m5bexnmwcc65dJgyBTp1qp3WqVOUnoU82DjnXDoUFcH06XDQQSBF36dPz8rJAeCz0ZxzLn2KirI2uNTlLRvnnHOxizXYSBotaYWkMknXJTneQdKscHyhpPyEY9eH9BWSRjWWp6TLQ5pJ2ichfYSkCkml4euG+GrsnHMumdi60STlAr8BRgJrgMWS5pjZOwmnXQB8amaHShoH3AqcI6kfMA7oD3QH5knqE66pL8+/As8ALyUpzgIzO63FK+mccy4lcbZshgJlZva+mW0DZgJj6pwzBng4vJ4NnCRJIX2mmX1pZh8AZSG/evM0szfNrDzG+jjnnGumOINND+DDhPdrQlrSc8ysEqgA9m7g2lTyTOYYSUslPSepf7ITJE2UVCKpZN26dSlk6ZxzLlW7wgSBN4CDzGwwMBV4KtlJZjbdzArNrLBbt26tWkDnnMt2cQabtcCBCe8PCGlJz5GUB3QBNjRwbSp51mJmm8xsS3g9F2iXOIHAOedc/OIMNouB3pJ6SWpPNOA/p845c4Dzw+uxwItmZiF9XJit1gvoDSxKMc9aJO0XxoGQNJSozhtapIbOtQG70Cr2LoPFNhvNzColXQ48D+QCD5jZMkk3AyVmNge4H3hUUhmwkSh4EM57DHgHqAQuM7MqiKY4180zpF8JXAPsB7wlaa6ZXUgUxC6VVAl8AYwLAc25rLeLrWLvMpj87+7XFRYWWklJSbqL4dxOy8+PAkxdBx0E5eWtXRqX7SQtMbPCZMd2hQkCzu2ydrFV7F0G82DjXBbbxVaxdxnMg41zWWwXW8XeZTAPNs5lsV1sFXuXwXyLAeey3C60ir3LYN6ycc45FzsPNs4552LnwcY551zsPNg455yLnQcb55xzsfPlapKQtA5IsshHvfYB1sdUnHTyerUtXq+2IxvrBNF2Lkn3aPFg0wIkldS3HlBb5vVqW7xebUc21qkx3o3mnHMudh5snHPOxc6DTcuYnu4CxMTr1bZ4vdqObKxTg3zMxjnnXOy8ZeOccy52Hmycc87FzoNNIGm0pBWSyiRdl+R4B0mzwvGFkvITjl0f0ldIGtVYnpLul7RU0luSZkvqnA31Sjh+l6Qt2VAnSQ9J+kBSafgqyJJ6SdIUSSsl/V3SlVlSrwUJv6uPJD2VJfU6SdIboV6vSDo0rnrFxsx2+S8gF3gPOBhoDywF+tU550fAPeH1OGBWeN0vnN8B6BXyyW0oT2DPhHzvAK7LhnqF6wqBR4Et2VAn4CFgbBb+G/wB8AiQE95/MxvqVSffx4HzsqFewErg8IR8H4r732RLf3nLJjIUKDOz981sGzATGFPnnDHAw+H1bOAkSQrpM83sSzP7ACgL+dWbp5ltgujTJdARiGuWRqvWS1IucBtwTUz1afU6taLWrtelwM1mtgPAzP4vS+oFgKQ9gX8F4mrZtHa9DNgzvO4CfBRTvWLjwSbSA/gw4f2akJb0HDOrBCqAvRu4tsE8JT0I/AM4DJjaEpVIorXrdTkwx8w+bqHyJ9PqvytgiqIuzzsldWiJSiTR2vU6BDhHUomk5yT1bqF61JWO3xfAd4H/rf5gF4PWrteFwFxJa4DvA7e0SC1akQebNDGzHwDdgb8D56S5ODtNUnfgLOILnOlyPdEHgqOAbwDXprc4LaYD8E+Llkz5HfBAmsvT0sYDM9JdiBZ0FXCKmR0APEjU/d6meLCJrAUOTHh/QEhLeo6kPKKm7IYGrm00TzOrImoqn7nTNUiuNet1BHAoUCapHOgkqaylKpKsvHXun/Scnf1dmdnHFvmS6D/50BarST1lrluGZOe0wL/BNcAT4fWTwKCdrkFyrf5/S9I+RL+nZ1ukBsm1Wr0kdQMGm9nCkD4LOLZlqtGK0j1olAlfQB7wPtFgXfXAXP8651xG7cG+x8Lr/tQe7HufaKAvaZ6AgEPDtQJuB25v6/VKcu+4Jgi0ap2A/RN+V78GbsmSet0C/DC8HgEszoZ6hesuAR6Ooz7pqFdIXw/0CddfADweZ/1i+ZmluwCZ8gWcQjTj4z1gcki7GTg9vN4N+CPRYN4i4OCEayeH61YA324kzxzgr8DfgLeBYhJmp7XVeiW5byzBprXrBLyY8Lv6PdA5S+rVleiT/9+A14g+Obf5eoVjLwGj46pPmn5fZ4Tf1dJQv4Pjrl9Lf/lyNc4552LnYzbOOedi58HGOedc7DzYOOeci50HG+ecc7HzYOOcc61M0lmSlknaIamwnnN2k7RI0aK9yyTdlHAs6UKqkvaS9GRY8WKRpAEh/UBJf5H0Tsjrx3XudYWk5eHYfzVS9gJJr4Vz35KU0kPpHmyca2GKccXrkP9cSV3D14+acf0ISc/EUTb3deHn/VCd5LeB7wHzG7j0S+BfzWwwUACMlnR0ODaB6AHQw8zscKKHwwF+BpSa2SDgPOC/Q3ol8FMz6wccDVwmqV8o34lEa7ANNrP+RM/+NWQr0QKn/YHRwK8ldW3kGg82zrU1ZnaKmX1G9KxMk4ONSz8z+7uZrWjkHDOz6g8u7cJX9bMq9S2k2o/o2TDMbDmQL2lfi1bCeCOkbyZaJqtHQl63WLRKRk1eknIl3SZpcWjBXByOrzSzd8Prj4D/A7o1VmcPNs61gtD18Hr4T/ukpL1C+kuSbg1dHislHR/SO0l6LHR7PBn2QykMx8rDkiy3AIeEPU5uq9tikXS3pAnh9ejQTfIG0Sfq6nN2l/RAuP+bklp7tWvXgPAHv5ToD/oL9tWSNfUtpLqU8PuVNBQ4iGjZm8Q884mWl6rOqw9wfPg39rKko0L6BUCFmR1FtDbgRZJ61clrKNFqB+81VhcPNs61jkeAa0P3xt+AXyYcyzOzocBPEtJ/BHwauj1+AQxJkud1wHtmVmBmk+q7saTdiBbb/E7IZ7+Ew5OBF8P9TwRuk7R7cyroagt/vEuB+4DT9dWmbqMau7aamVWZWQFRwBhaPQZD/Qup3gJ0Dfe9AngTqEooU2eifX5+Yl+tiJ1HtMjs0cAk4DFJAk4Gzgt5LSRasbp3Ql77E+1d9YPqFlZD8lKttHOueSR1Abqa2csh6WGiZUyqVS+IuQTID6//hdDfbmZvS3prJ4pwGPBBddeHpN8DE8Oxk4n+EF4d3u8G9CTqZnE7wcyGQTRmA0wwswk7kddnkv5CNEbyNl9fSPXBcN4moo3xqvfL+oBovTUktSMKNMVm9kRC9muAJyxaTmaRpB3APkTrAV5hZs/XLY+i/YKeJVpS5/VU6uAtG+fS78vwvYqd+wBYSe3/07ulcI2AM0PrqMDMepqZB5oMIKlb9cC7pI7ASGB5OPwUUUsUYDjRemqESSPtQ/qFwHwz2xQCz/3A382s7vYENXlJ6kPULbYeeB64NAQpJPUJ3a7tiQLcI2Y2O9X6eLBxLmZmVgF8Wj0eQ7T51csNXALRYq1nA4RZQwOTnLMZ2CPh/Sqgn6QO4Y/USSG9eqD4kPB+fMI1zwNXhD9GSDoitVq5nSHpDEUboR0DPCvp+ZDeXdLccNr+wF9Cq3Yx0ZhN9ZjcLcCZkv4G/CdRYAE4HHhb0grg20D1FOfjiP7d/WtCd94p4dgDwMGS3iaa1XZ+aOXcB7wDvBGO3Uv0Yehs4ARgQkJeBY3W2RfidK5lhW6IxG177yCaIXQP0ImoW+MHZvappJeAq82sJAz6l5hZfhg3eZhodtFyon3pzzKzdxXtF1RoZusl/YFoL5rnzGySomckziDqPtlCtHPqQ5JGE22RsBVYABxiZqeFT8y/JtofJYeou+20GH88bhflwca5DCQpF2hnZv8MLZJ5QF+L9qZ3rs3xCQLOZaZORF0o7YjGVX7kgca1Zd6ycc45FzufIOCccy52Hmycc87FzoONc8652Hmwcc45FzsPNs4552L3/wFYLJVbqQFwCAAAAABJRU5ErkJggg==\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "uJam-KLXq9TY"
      },
      "source": [
        "##Models\n",
        "\n",
        "At first, we create a Baseline Model using mathematical relationship between distance and RSSI. This model and its results will serve as reference for other ML model's performance\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "GxInTL2hia9s"
      },
      "source": [
        "### Baseline Model"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "JnWDVno7HlYE"
      },
      "source": [
        "test_semi['Baseline Dist'] = 10.43*(10**((-61-test_semi['RSSI(dBm)'])/27))\n",
        "test_semi['Baseline Dist'].values[test_semi['Baseline Dist'].values>test_semi['Distance'].max()]=test_semi['Distance'].max()"
      ],
      "execution_count": 29,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "mYpk21Uwl2Ze"
      },
      "source": [
        "Results of Baseline Model:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "fWixpeOyIZb_",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "eead646a-d2b5-4c8a-8f90-83a441c30747"
      },
      "source": [
        "dist_err=((test_semi['Distance']-test_semi['Baseline Dist'])**2)**0.5\n",
        "df_error = pd.DataFrame(dist_err, columns = ['Distance Error'])\n",
        "print(\"MEAN DISTANCE ERROR USING BASELINE MODEL :\", dist_err.mean(),'m')\n",
        "print(\"MAXIMUM DISTANCE ERROR USING BASELINE MODEL :\", dist_err.max(),'m')"
      ],
      "execution_count": 30,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "MEAN DISTANCE ERROR USING BASELINE MODEL : 12.131676973077862 m\n",
            "MAXIMUM DISTANCE ERROR USING BASELINE MODEL : 32.460504981009464 m\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "t5vl-5QHvEDH",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "e2817da9-c8f2-4981-aae2-e9740290321e"
      },
      "source": [
        "count = len([i for i in dist_err if i  < 2])\n",
        "low_error = count / len(dist_err) * 100\n",
        "print(\"Percentage of data with less than 2m error : \",low_error,\"%\")"
      ],
      "execution_count": 31,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Percentage of data with less than 2m error :  7.629427792915531 %\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Rr8cMwbtIlYq",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 295
        },
        "outputId": "d122007d-5ae7-430f-badb-4827c0f8300a"
      },
      "source": [
        "x_ax = range(len(test_semi['Baseline Dist']))\n",
        "plt.plot(x_ax, test_semi['Distance'], linewidth=1, label=\"original\")\n",
        "plt.plot(x_ax, test_semi['Baseline Dist'], linewidth=1.1, label=\"predicted\")\n",
        "plt.title(\"Baseline Prediction\")\n",
        "plt.xlabel('Observation Number')\n",
        "plt.ylabel('Distance')\n",
        "plt.legend(loc='best',fancybox=True, shadow=True)\n",
        "# plt.grid(True)\n",
        "plt.show() "
      ],
      "execution_count": 32,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "LxBVYXqTdN7x"
      },
      "source": [
        "For Machine Learning model development we follow two main concepts\n",
        "1. RSSI Distance Estimation\n",
        "2. RSSI Fingerprinting\n",
        "\n",
        "### 1. RSSI Distance Estimation (Multilateration)\n",
        "\n",
        "RSSI distance estimation is solely dependent on a clear correlation between distance and RSSI value. As mentioned previously clear correlation is not present, thereby giving us considerable distance error between the predicted location and actual location.\n",
        "\n",
        "For performing the same we have tried several algorithms, the following have yielded the best results:"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "3ytogEDQ0hU9"
      },
      "source": [
        "### Linear Regression using Polynomial Features of Degree 2:\n",
        "\n",
        "\n",
        "Since from the formula of RSSI and distance :\n",
        "\n",
        "![picture](https://drive.google.com/uc?export=view&id=15tuh-QXcKEVx1P8EW9MuWRd_lJlTpywg)\n",
        "\n",
        "where n=2, an inverse square law exists between RSSI and distance, thereby motivating us to use the same.\n",
        "\n",
        "The following illustrate the results obtained:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "jHGCX79WSEB4",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "5be4b823-d0f4-4129-a55f-493e4d1fdd7b"
      },
      "source": [
        "trainX_final = train_semi[['Latitude_AP', 'Longitude_AP', 'Channel', 'RSSI(dBm)', 'Hight Difference(m)', 'Outdoor', 'Moving Average (RSSI)', 1, 2, 3, 4]]\n",
        "trainY_final = train_semi[['Latitude_OP','Longitude_OP','Distance']]\n",
        "testX_final = test_semi[['Latitude_AP', 'Longitude_AP', 'Channel', 'RSSI(dBm)', 'Hight Difference(m)', 'Outdoor', 'Moving Average (RSSI)', 1, 2, 3, 4]]\n",
        "testY_final = test_semi[['Latitude_OP','Longitude_OP','Distance']]\n",
        "mm_scaler = MinMaxScaler()\n",
        "train_X = pd.DataFrame(mm_scaler.fit_transform(trainX_final),columns=['Latitude_AP', 'Longitude_AP', 'Channel', 'RSSI(dBm)', 'Hight Difference(m)', 'Outdoor', 'Moving Average (RSSI)', '1', '2', '3', '4'])\n",
        "test_X = pd.DataFrame(mm_scaler.transform(testX_final),columns=['Latitude_AP', 'Longitude_AP', 'Channel', 'RSSI(dBm)', 'Hight Difference(m)', 'Outdoor', 'Moving Average (RSSI)', '1', '2', '3', '4'])"
      ],
      "execution_count": 33,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.7/dist-packages/sklearn/utils/validation.py:1679: FutureWarning: Feature names only support names that are all strings. Got feature names with dtypes: ['int', 'str']. An error will be raised in 1.2.\n",
            "  FutureWarning,\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/utils/validation.py:1679: FutureWarning: Feature names only support names that are all strings. Got feature names with dtypes: ['int', 'str']. An error will be raised in 1.2.\n",
            "  FutureWarning,\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/utils/validation.py:1679: FutureWarning: Feature names only support names that are all strings. Got feature names with dtypes: ['int', 'str']. An error will be raised in 1.2.\n",
            "  FutureWarning,\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "2QH5SEwctitd"
      },
      "source": [
        "####Estimation of Distance of RP from AP using Linear Regression with Polynomial Feature of Degree 2"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "kqvyVi92VUnU",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 0
        },
        "outputId": "0bcbd0c6-3b81-4353-9c7b-58fc3943a998"
      },
      "source": [
        "from sklearn.linear_model import LinearRegression\n",
        "from sklearn.preprocessing import PolynomialFeatures\n",
        "\n",
        "polyfeat = PolynomialFeatures(degree=2)\n",
        "newtrain_X = polyfeat.fit_transform(train_X[['Channel', 'Hight Difference(m)', 'Outdoor', 'Moving Average (RSSI)']])\n",
        "newtest_X = polyfeat.transform(test_X[['Channel', 'Hight Difference(m)', 'Outdoor', 'Moving Average (RSSI)']])\n",
        "\n",
        "model1 = LinearRegression(normalize=True, fit_intercept=False)\n",
        "model1.fit(newtrain_X, trainY_final['Distance'])\n",
        "pred_y = model1.predict(newtest_X)\n",
        "\n",
        "dist_err=((testY_final['Distance']-pred_y)**2)**0.5\n",
        "df_error = pd.DataFrame(dist_err)\n",
        "df_error.clip(upper=pd.Series({'Distance': 38}), axis=1, inplace=True)\n",
        "\n",
        "plt.plot(testY_final['Distance'], linewidth=1, label=\"original\")\n",
        "plt.plot(pred_y, linewidth=1, label=\"predicted\")\n",
        "plt.title(\"Orginal vs Predicted distance for Polynomial Regression (Deg=2)\")\n",
        "plt.xlabel('Observation Number')\n",
        "plt.ylabel('Distance')\n",
        "plt.legend(loc='best',fancybox=True, shadow=True)\n",
        "plt.show() "
      ],
      "execution_count": 34,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.7/dist-packages/sklearn/linear_model/_base.py:145: FutureWarning: 'normalize' was deprecated in version 1.0 and will be removed in 1.2.\n",
            "If you wish to scale the data, use Pipeline with a StandardScaler in a preprocessing stage. To reproduce the previous behavior:\n",
            "\n",
            "from sklearn.pipeline import make_pipeline\n",
            "\n",
            "model = make_pipeline(StandardScaler(with_mean=False), LinearRegression())\n",
            "\n",
            "If you wish to pass a sample_weight parameter, you need to pass it as a fit parameter to each step of the pipeline as follows:\n",
            "\n",
            "kwargs = {s[0] + '__sample_weight': sample_weight for s in model.steps}\n",
            "model.fit(X, y, **kwargs)\n",
            "\n",
            "\n",
            "  FutureWarning,\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "SjGkMn-hl7g5"
      },
      "source": [
        "Results obtained are as follows:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "R3Gpxz8ThLxX",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "2d2fb9ee-8afa-4f05-a023-31c56a7435be"
      },
      "source": [
        "print(\"MEAN DISTANCE ERROR USING POLYNOMIAL REGRESSOR (DEGREE 2): \", df_error['Distance'].mean(),\"m\")\n",
        "print(\"MAXIMUM DISTANCE ERROR USING POLYNOMIAL REGRESSOR(DEGREE 2): \", df_error['Distance'].max(),\"m\")"
      ],
      "execution_count": 35,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "MEAN DISTANCE ERROR USING POLYNOMIAL REGRESSOR (DEGREE 2):  10.0736354919281 m\n",
            "MAXIMUM DISTANCE ERROR USING POLYNOMIAL REGRESSOR(DEGREE 2):  25.791606937720367 m\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "YFgo-BA7woDG",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "e3bea89a-3e40-444e-cc63-b6550ec3a663"
      },
      "source": [
        "count = len([i for i in df_error['Distance'] if i  < 2])\n",
        "low_error = count / len(df_error['Distance']) * 100\n",
        "print(\"Percentage of data with less than 2m error : \",low_error,\"%\")"
      ],
      "execution_count": 36,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Percentage of data with less than 2m error :  11.807447774750226 %\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "RZxgP9dTtbf7"
      },
      "source": [
        "####Estimation of Latitude and Longitude using Linear Regression with Polynomial Feature of Degree 2"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "VTTp3QNYahMB",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 538
        },
        "outputId": "b011a41d-4187-4574-e2ff-dcede57e9fc2"
      },
      "source": [
        "polyfeat = PolynomialFeatures(degree=2)\n",
        "newtrain_X = polyfeat.fit_transform(train_X[['Latitude_AP','Longitude_AP','Channel', 'Hight Difference(m)', 'Outdoor', 'Moving Average (RSSI)','1','2','3','4']])\n",
        "newtest_X = polyfeat.transform(test_X[['Latitude_AP','Longitude_AP','Channel', 'Hight Difference(m)', 'Outdoor', 'Moving Average (RSSI)','1','2','3','4']])\n",
        "\n",
        "model2 = LinearRegression(normalize=True, fit_intercept=False)\n",
        "model2.fit(newtrain_X, trainY_final[['Latitude_OP','Longitude_OP']])\n",
        "pred_y = model2.predict(newtest_X)\n",
        "\n",
        "pred = pd.DataFrame(pred_y, columns = ['Latitude','Longitude'])\n",
        "pred['dist_err'] = (((pred['Latitude']-testY_final['Latitude_OP'])*111000)**2 + (((pred['Longitude']-testY_final['Longitude_OP'])*111000)**2))**0.5\n",
        "pred.clip(upper=pd.Series({'dist_err': 38}), axis=1,inplace=True)\n",
        "\n",
        "plt.plot(testY_final[['Latitude_OP']],label=\"original\")\n",
        "plt.plot(pred_y[:,0],label=\"predicted\")\n",
        "plt.title(\"Latitude original vs predicted\")\n",
        "plt.xlabel(\"Observation Number\")\n",
        "plt.ylabel(\"Latitude\")\n",
        "plt.legend(loc='best',fancybox=True, shadow=True)\n",
        "plt.show()"
      ],
      "execution_count": 37,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.7/dist-packages/sklearn/linear_model/_base.py:145: FutureWarning: 'normalize' was deprecated in version 1.0 and will be removed in 1.2.\n",
            "If you wish to scale the data, use Pipeline with a StandardScaler in a preprocessing stage. To reproduce the previous behavior:\n",
            "\n",
            "from sklearn.pipeline import make_pipeline\n",
            "\n",
            "model = make_pipeline(StandardScaler(with_mean=False), LinearRegression())\n",
            "\n",
            "If you wish to pass a sample_weight parameter, you need to pass it as a fit parameter to each step of the pipeline as follows:\n",
            "\n",
            "kwargs = {s[0] + '__sample_weight': sample_weight for s in model.steps}\n",
            "model.fit(X, y, **kwargs)\n",
            "\n",
            "\n",
            "  FutureWarning,\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "k2P96T2tbGvN",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 295
        },
        "outputId": "ba57ee96-125b-42c7-ea1f-24ff44a1e1cb"
      },
      "source": [
        "plt.plot(testY_final[['Longitude_OP']],label=\"original\")\n",
        "plt.plot(pred_y[:,1],label=\"predicted\")\n",
        "plt.title(\"Longitude original vs predicted\")\n",
        "plt.xlabel(\"Observation Number\")\n",
        "plt.ylabel(\"Longitude\")\n",
        "plt.legend(loc='best',fancybox=True, shadow=True)\n",
        "plt.show()"
      ],
      "execution_count": 38,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "LuSmfQJlmBNy"
      },
      "source": [
        "Results Obtained"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "zIr9bwMDg1ns",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "723466d0-f9f6-4273-a636-a512b83175f1"
      },
      "source": [
        "print(\"MEAN DISTANCE ERROR USING POLYNOMIAL REGRESSOR (DEGREE 2): \", pred['dist_err'].mean(),\"m\")\n",
        "print(\"MAXIMUM DISTANCE ERROR USING POLYNOMIAL REGRESSOR(DEGREE 2): \", pred['dist_err'].max(),\"m\")"
      ],
      "execution_count": 39,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "MEAN DISTANCE ERROR USING POLYNOMIAL REGRESSOR (DEGREE 2):  22.529960865882906 m\n",
            "MAXIMUM DISTANCE ERROR USING POLYNOMIAL REGRESSOR(DEGREE 2):  38.0 m\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "liT1sOHNwyi-",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "ea582e0a-efd6-48c9-8aff-d3e7fe7bbc58"
      },
      "source": [
        "count = len([i for i in pred['dist_err'] if i  < 2])\n",
        "low_error = count / len(pred['dist_err']) * 100\n",
        "print(\"Percentage of data with less than 2m error : \",low_error,\"%\")"
      ],
      "execution_count": 40,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Percentage of data with less than 2m error :  0.0 %\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "2mBhtd4krRxh"
      },
      "source": [
        "###AdaBoost\n",
        "\n",
        "As Adaptive Boosting follows an ensemble approach it significantly improves the results from Linear Regressor, thus motivating us to put it to use.\n",
        "\n",
        "We created 4 models each for each AP estimating the distance of the RP from the AP."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Defq5TlStJ8N"
      },
      "source": [
        "####Distance Estimation using AdaBoost"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Npf0Q0fzApIe",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "8d2c6aac-470c-42fd-c347-5e7ad917a0c4"
      },
      "source": [
        "from sklearn.ensemble import AdaBoostRegressor\n",
        "from sklearn.model_selection import cross_val_score, KFold\n",
        "from sklearn.metrics import mean_squared_error\n",
        "\n",
        "mm_scaler = MinMaxScaler()\n",
        "train_semiX = pd.DataFrame(mm_scaler.fit_transform(train_semi[['Latitude_AP', 'Longitude_AP', 'Channel', 'RSSI(dBm)', 'Hight Difference(m)', 'Outdoor', 'Moving Average (RSSI)', 1, 2, 3, 4]]),columns=['Latitude_AP', 'Longitude_AP', 'Channel', 'RSSI(dBm)', 'Hight Difference(m)', 'Outdoor', 'Moving Average (RSSI)', 1, 2, 3, 4])\n",
        "test_semiX = pd.DataFrame(mm_scaler.transform(test_semi[['Latitude_AP', 'Longitude_AP', 'Channel', 'RSSI(dBm)', 'Hight Difference(m)', 'Outdoor', 'Moving Average (RSSI)', 1, 2, 3, 4]]),columns=['Latitude_AP', 'Longitude_AP', 'Channel', 'RSSI(dBm)', 'Hight Difference(m)', 'Outdoor', 'Moving Average (RSSI)', 1, 2, 3, 4])\n",
        "\n",
        "ap01_dat_tr = pd.concat([train_semiX[train_semiX[1]==1],(train_semi[train_semi[1]==1])[['Distance','Latitude_OP','Longitude_OP']]],axis=1).reset_index(drop=True)\n",
        "ap02_dat_tr = pd.concat([train_semiX[train_semiX[2]==1],(train_semi[train_semi[2]==1])[['Distance','Latitude_OP','Longitude_OP']]],axis=1).reset_index(drop=True)\n",
        "ap03_dat_tr = pd.concat([train_semiX[train_semiX[3]==1],(train_semi[train_semi[3]==1])[['Distance','Latitude_OP','Longitude_OP']]],axis=1).reset_index(drop=True)\n",
        "ap04_dat_tr = pd.concat([train_semiX[train_semiX[4]==1],(train_semi[train_semi[4]==1])[['Distance','Latitude_OP','Longitude_OP']]],axis=1).reset_index(drop=True)\n",
        "ap01_dat_te = pd.concat([test_semiX[test_semiX[1]==1],(test_semi[test_semi[1]==1])[['Distance','Latitude_OP','Longitude_OP']]],axis=1).reset_index(drop=True)\n",
        "ap02_dat_te = pd.concat([test_semiX[test_semiX[2]==1],(test_semi[test_semi[2]==1])[['Distance','Latitude_OP','Longitude_OP']]],axis=1).reset_index(drop=True)\n",
        "ap03_dat_te = pd.concat([test_semiX[test_semiX[3]==1],(test_semi[test_semi[3]==1])[['Distance','Latitude_OP','Longitude_OP']]],axis=1).reset_index(drop=True)\n",
        "ap04_dat_te = pd.concat([test_semiX[test_semiX[4]==1],(test_semi[test_semi[4]==1])[['Distance','Latitude_OP','Longitude_OP']]],axis=1).reset_index(drop=True)\n",
        "ap01_datX_tr = ap01_dat_tr[['Outdoor','Moving Average (RSSI)']]\n",
        "ap01_datY_tr = ap01_dat_tr[['Distance']]\n",
        "ap02_datX_tr = ap02_dat_tr[['Outdoor','Moving Average (RSSI)']]\n",
        "ap02_datY_tr = ap02_dat_tr[['Distance']]\n",
        "ap03_datX_tr = ap03_dat_tr[['Outdoor','Moving Average (RSSI)']]\n",
        "ap03_datY_tr = ap03_dat_tr[['Distance']]\n",
        "ap04_datX_tr = ap04_dat_tr[['Outdoor','Moving Average (RSSI)']]\n",
        "ap04_datY_tr = ap04_dat_tr[['Distance']]\n",
        "ap01_datX_te = ap01_dat_te[['Outdoor','Moving Average (RSSI)']]\n",
        "ap01_datY_te = ap01_dat_te[['Distance']]\n",
        "ap02_datX_te = ap02_dat_te[['Outdoor','Moving Average (RSSI)']]\n",
        "ap02_datY_te = ap02_dat_te[['Distance']]\n",
        "ap03_datX_te = ap03_dat_te[['Outdoor','Moving Average (RSSI)']]\n",
        "ap03_datY_te = ap03_dat_te[['Distance']]\n",
        "ap04_datX_te = ap04_dat_te[['Outdoor','Moving Average (RSSI)']]\n",
        "ap04_datY_te = ap04_dat_te[['Distance']]\n",
        "polyfeat = PolynomialFeatures(degree=2)\n",
        "ap01_datX_tr = polyfeat.fit_transform(ap01_datX_tr)\n",
        "ap01_datX_te = polyfeat.transform(ap01_datX_te)\n",
        "polyfeat = PolynomialFeatures(degree=2)\n",
        "ap02_datX_tr = polyfeat.fit_transform(ap02_datX_tr)\n",
        "ap02_datX_te = polyfeat.transform(ap02_datX_te)\n",
        "polyfeat = PolynomialFeatures(degree=2)\n",
        "ap03_datX_tr = polyfeat.fit_transform(ap03_datX_tr)\n",
        "ap03_datX_te = polyfeat.transform(ap03_datX_te)\n",
        "polyfeat = PolynomialFeatures(degree=2)\n",
        "ap04_datX_tr = polyfeat.fit_transform(ap04_datX_tr)\n",
        "ap04_datX_te = polyfeat.transform(ap04_datX_te)\n",
        "\n",
        "ada_AP1 = AdaBoostRegressor(n_estimators=50)\n",
        "ada_AP1.fit(ap01_datX_tr,ap01_datY_tr)\n",
        "prediction = ada_AP1.predict(ap01_datX_te)\n",
        "prediction = prediction.reshape(-1,1)\n",
        "\n",
        "total_dist_err = []\n",
        "dist=[]\n",
        "dist=dist+list(prediction[:,0])\n",
        "\n",
        "dist_err=((ap01_datY_te-prediction)**2)**0.5\n",
        "df_error = pd.DataFrame(dist_err, columns = ['Distance Error'])\n",
        "dist_err.clip(upper=pd.Series({'Distance': 38}), axis=1,inplace=True)\n",
        "total_dist_err += list(dist_err.values)\n",
        "\n",
        "ada_AP2 = AdaBoostRegressor(n_estimators=50)\n",
        "ada_AP2.fit(ap02_datX_tr,ap02_datY_tr)\n",
        "prediction = ada_AP2.predict(ap02_datX_te)\n",
        "prediction = prediction.reshape(-1,1)\n",
        "\n",
        "dist=dist+list(prediction)\n",
        "dist_err=((ap02_datY_te-prediction)**2)**0.5\n",
        "df_error = pd.DataFrame(dist_err, columns = ['Distance Error'])\n",
        "dist_err.clip(upper=pd.Series({'Distance': 38}), axis=1,inplace=True)\n",
        "total_dist_err += list(dist_err.values)\n",
        "\n",
        "ada_AP3 = AdaBoostRegressor(n_estimators=50)\n",
        "ada_AP3.fit(ap03_datX_tr,ap03_datY_tr)\n",
        "prediction = ada_AP3.predict(ap03_datX_te)\n",
        "prediction = prediction.reshape(-1,1)\n",
        "\n",
        "dist=dist+list(prediction[:,0])\n",
        "dist_err=((ap03_datY_te-prediction)**2)**0.5\n",
        "df_error = pd.DataFrame(dist_err, columns = ['Distance Error'])\n",
        "dist_err.clip(upper=pd.Series({'Distance': 38}), axis=1,inplace=True)\n",
        "total_dist_err += list(dist_err.values)\n",
        "\n",
        "ada_AP4 = AdaBoostRegressor(n_estimators=50)\n",
        "ada_AP4.fit(ap04_datX_tr,ap04_datY_tr)\n",
        "prediction = ada_AP4.predict(ap04_datX_te)\n",
        "prediction = prediction.reshape(-1,1)\n",
        "\n",
        "dist=dist+list(prediction[:,0])\n",
        "dist_err=((ap04_datY_te-prediction)**2)**0.5\n",
        "df_error = pd.DataFrame(dist_err, columns = ['Distance Error'])\n",
        "dist_err.clip(upper=pd.Series({'Distance': 38}), axis=1,inplace=True)\n",
        "total_dist_err += list(dist_err.values)"
      ],
      "execution_count": 41,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.7/dist-packages/sklearn/utils/validation.py:1679: FutureWarning: Feature names only support names that are all strings. Got feature names with dtypes: ['int', 'str']. An error will be raised in 1.2.\n",
            "  FutureWarning,\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/utils/validation.py:1679: FutureWarning: Feature names only support names that are all strings. Got feature names with dtypes: ['int', 'str']. An error will be raised in 1.2.\n",
            "  FutureWarning,\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/utils/validation.py:1679: FutureWarning: Feature names only support names that are all strings. Got feature names with dtypes: ['int', 'str']. An error will be raised in 1.2.\n",
            "  FutureWarning,\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/utils/validation.py:985: DataConversionWarning: A column-vector y was passed when a 1d array was expected. Please change the shape of y to (n_samples, ), for example using ravel().\n",
            "  y = column_or_1d(y, warn=True)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/utils/validation.py:985: DataConversionWarning: A column-vector y was passed when a 1d array was expected. Please change the shape of y to (n_samples, ), for example using ravel().\n",
            "  y = column_or_1d(y, warn=True)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/utils/validation.py:985: DataConversionWarning: A column-vector y was passed when a 1d array was expected. Please change the shape of y to (n_samples, ), for example using ravel().\n",
            "  y = column_or_1d(y, warn=True)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/utils/validation.py:985: DataConversionWarning: A column-vector y was passed when a 1d array was expected. Please change the shape of y to (n_samples, ), for example using ravel().\n",
            "  y = column_or_1d(y, warn=True)\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "TjtzF--CmNn4"
      },
      "source": [
        "Results obtained"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "esOElQT9Q5WW",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "b7762b2d-cac6-4ac0-9076-7d5baedd8dad"
      },
      "source": [
        "adamean=sum(total_dist_err)/len(total_dist_err)\n",
        "adamax=max(total_dist_err)\n",
        "print(\"MEAN DISTANCE ERROR USING ADABOOST REGRESSOR: \", adamean[0],'m')\n",
        "print(\"MAXIMUM DISTANCE ERROR USING ADABOOST REGRESSOR: \",adamax[0],'m')"
      ],
      "execution_count": 42,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "MEAN DISTANCE ERROR USING ADABOOST REGRESSOR:  14.175428476665237 m\n",
            "MAXIMUM DISTANCE ERROR USING ADABOOST REGRESSOR:  34.413591805764185 m\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "UQobP4r_u2_B",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "df513708-7887-412f-bc1b-df8befbfeb5b"
      },
      "source": [
        "count = len([i for i in total_dist_err if i  < 2])\n",
        "low_error = count / len(total_dist_err) * 100\n",
        "print(\"Percentage of data with less than 2m error : \",low_error,'%')"
      ],
      "execution_count": 43,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Percentage of data with less than 2m error :  5.812897366030881 %\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "1vzgdwjKbK77",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 350
        },
        "outputId": "6412c6f1-67f1-4c5a-dee5-9b931302e52c"
      },
      "source": [
        "x_ax = range(len(dist))\n",
        "plt.plot(x_ax, testY_final['Distance'], linewidth=1, label=\"original\")\n",
        "plt.plot(x_ax, dist, linewidth=1, label=\"predicted\")\n",
        "plt.title(\"Orginal vs Predicted distance for Polynomial Regression (Deg=2)\")\n",
        "plt.ylabel('Distance')\n",
        "plt.xlabel('Observation Number')\n",
        "plt.legend(loc='best',fancybox=True, shadow=True)\n",
        "# plt.grid(True)\n",
        "plt.show() "
      ],
      "execution_count": 44,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.7/dist-packages/numpy/core/_asarray.py:136: VisibleDeprecationWarning: Creating an ndarray from ragged nested sequences (which is a list-or-tuple of lists-or-tuples-or ndarrays with different lengths or shapes) is deprecated. If you meant to do this, you must specify 'dtype=object' when creating the ndarray\n",
            "  return array(a, dtype, copy=False, order=order, subok=True)\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "9FBMb7hCtRA8"
      },
      "source": [
        "####Direct estimation of Latitude and Longitude using AdaBoost\n",
        "\n",
        "In this case we trained 8 different models for the 4 different APs.\n",
        "Two models for each AP for predicting Latitude and Longitude of the reference points/observation points respectively."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "t6eBqLnnoIbX"
      },
      "source": [
        "ap01_datX_tr = ap01_dat_tr[['Latitude_AP','Longitude_AP','Outdoor','Moving Average (RSSI)']]\n",
        "ap01_datY_tr = ap01_dat_tr[['Latitude_OP','Longitude_OP']]\n",
        "ap02_datX_tr = ap02_dat_tr[['Latitude_AP','Longitude_AP','Outdoor','Moving Average (RSSI)']]\n",
        "ap02_datY_tr = ap02_dat_tr[['Latitude_OP','Longitude_OP']]\n",
        "ap03_datX_tr = ap03_dat_tr[['Latitude_AP','Longitude_AP','Outdoor','Moving Average (RSSI)']]\n",
        "ap03_datY_tr = ap03_dat_tr[['Latitude_OP','Longitude_OP']]\n",
        "ap04_datX_tr = ap04_dat_tr[['Latitude_AP','Longitude_AP','Outdoor','Moving Average (RSSI)']]\n",
        "ap04_datY_tr = ap04_dat_tr[['Latitude_OP','Longitude_OP']]\n",
        "ap01_datX_te = ap01_dat_te[['Latitude_AP','Longitude_AP','Outdoor','Moving Average (RSSI)']]\n",
        "ap01_datY_te = ap01_dat_te[['Latitude_OP','Longitude_OP']]\n",
        "ap02_datX_te = ap02_dat_te[['Latitude_AP','Longitude_AP','Outdoor','Moving Average (RSSI)']]\n",
        "ap02_datY_te = ap02_dat_te[['Latitude_OP','Longitude_OP']]\n",
        "ap03_datX_te = ap03_dat_te[['Latitude_AP','Longitude_AP','Outdoor','Moving Average (RSSI)']]\n",
        "ap03_datY_te = ap03_dat_te[['Latitude_OP','Longitude_OP']]\n",
        "ap04_datX_te = ap04_dat_te[['Latitude_AP','Longitude_AP','Outdoor','Moving Average (RSSI)']]\n",
        "ap04_datY_te = ap04_dat_te[['Latitude_OP','Longitude_OP']]\n",
        "\n",
        "ada_AP1_Lat = AdaBoostRegressor(n_estimators=25)\n",
        "ada_AP1_Lat.fit(ap01_datX_tr,ap01_datY_tr['Latitude_OP'])\n",
        "prediction1_Lat = ada_AP1_Lat.predict(ap01_datX_te)\n",
        "prediction1_Lat = pd.DataFrame(prediction1_Lat.reshape(-1,1),columns=['Latitude'])\n",
        "ada_AP1_Lon = AdaBoostRegressor(n_estimators=25)\n",
        "ada_AP1_Lon.fit(ap01_datX_tr,ap01_datY_tr['Longitude_OP'])\n",
        "prediction1_Lon = ada_AP1_Lon.predict(ap01_datX_te)\n",
        "prediction1_Lon = pd.DataFrame(prediction1_Lon.reshape(-1,1),columns=['Longitude'])\n",
        "\n",
        "total_dist_err = []\n",
        "pred1 = pd.concat([prediction1_Lat,prediction1_Lon],axis=1)\n",
        "pred['dist_err'] = (((pred1['Latitude']-ap01_datY_tr['Latitude_OP'])*111000)**2 + (((pred1['Longitude']-ap01_datY_tr['Longitude_OP'])*111000)**2))**0.5\n",
        "pred.clip(upper=pd.Series({'dist_err': 38}), axis=1,inplace=True)\n",
        "total_dist_err += list(dist_err.values)\n",
        "\n",
        "ada_AP2_Lat = AdaBoostRegressor(n_estimators=25)\n",
        "ada_AP2_Lat.fit(ap02_datX_tr,ap02_datY_tr['Latitude_OP'])\n",
        "prediction2_Lat = ada_AP2_Lat.predict(ap02_datX_te)\n",
        "prediction2_Lat = pd.DataFrame(prediction2_Lat.reshape(-1,1),columns=['Latitude'])\n",
        "ada_AP2_Lon = AdaBoostRegressor(n_estimators=25)\n",
        "ada_AP2_Lon.fit(ap02_datX_tr,ap01_datY_tr['Longitude_OP'])\n",
        "prediction2_Lon = ada_AP2_Lon.predict(ap02_datX_te)\n",
        "prediction2_Lon = pd.DataFrame(prediction2_Lon.reshape(-1,1),columns=['Longitude'])\n",
        "\n",
        "pred2 = pd.concat([prediction2_Lat,prediction2_Lon],axis=1)\n",
        "pred['dist_err'] = (((pred2['Latitude']-ap02_datY_tr['Latitude_OP'])*111000)**2 + (((pred2['Longitude']-ap02_datY_tr['Longitude_OP'])*111000)**2))**0.5\n",
        "pred.clip(upper=pd.Series({'dist_err': 38}), axis=1,inplace=True)\n",
        "total_dist_err += list(dist_err.values)\n",
        "\n",
        "ada_AP3_Lat = AdaBoostRegressor(n_estimators=25)\n",
        "ada_AP3_Lat.fit(ap03_datX_tr,ap03_datY_tr['Latitude_OP'])\n",
        "prediction3_Lat = ada_AP3_Lat.predict(ap03_datX_te)\n",
        "prediction3_Lat = pd.DataFrame(prediction3_Lat.reshape(-1,1),columns=['Latitude'])\n",
        "ada_AP3_Lon = AdaBoostRegressor(n_estimators=25)\n",
        "ada_AP3_Lon.fit(ap03_datX_tr,ap03_datY_tr['Longitude_OP'])\n",
        "prediction3_Lon = ada_AP3_Lon.predict(ap03_datX_te)\n",
        "prediction3_Lon = pd.DataFrame(prediction3_Lon.reshape(-1,1),columns=['Longitude'])\n",
        "\n",
        "pred3 = pd.concat([prediction3_Lat,prediction3_Lon],axis=1)\n",
        "pred['dist_err'] = (((pred3['Latitude']-ap03_datY_tr['Latitude_OP'])*111000)**2 + (((pred3['Longitude']-ap03_datY_tr['Longitude_OP'])*111000)**2))**0.5\n",
        "pred.clip(upper=pd.Series({'dist_err': 38}), axis=1,inplace=True)\n",
        "total_dist_err += list(dist_err.values)\n",
        "\n",
        "ada_AP4_Lat = AdaBoostRegressor(n_estimators=25)\n",
        "ada_AP4_Lat.fit(ap04_datX_tr,ap04_datY_tr['Latitude_OP'])\n",
        "prediction4_Lat = ada_AP4_Lat.predict(ap04_datX_te)\n",
        "prediction4_Lat = pd.DataFrame(prediction4_Lat.reshape(-1,1),columns=['Latitude'])\n",
        "ada_AP4_Lon = AdaBoostRegressor(n_estimators=25)\n",
        "ada_AP4_Lon.fit(ap04_datX_tr,ap04_datY_tr['Longitude_OP'])\n",
        "prediction4_Lon = ada_AP4_Lon.predict(ap04_datX_te)\n",
        "prediction4_Lon = pd.DataFrame(prediction4_Lon.reshape(-1,1),columns=['Longitude'])\n",
        "\n",
        "pred4 = pd.concat([prediction4_Lat,prediction4_Lon],axis=1)\n",
        "pred['dist_err'] = (((pred4['Latitude']-ap04_datY_tr['Latitude_OP'])*111000)**2 + (((pred4['Longitude']-ap04_datY_tr['Longitude_OP'])*111000)**2))**0.5\n",
        "pred.clip(upper=pd.Series({'dist_err': 38}), axis=1,inplace=True)\n",
        "total_dist_err += list(dist_err.values)"
      ],
      "execution_count": 45,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "sck4QtLkkpZL"
      },
      "source": [
        "Results obtained"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "1GQmcI_Oe0_5",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "0b86c840-99f6-4507-e910-8e7ef65d516a"
      },
      "source": [
        "adamean=sum(total_dist_err)/len(total_dist_err)\n",
        "adamax=max(total_dist_err)\n",
        "print(\"MEAN DISTANCE ERROR USING ADABOOST REGRESSOR: \", adamean[0],'m')\n",
        "print(\"MAXIMUM DISTANCE ERROR USING ADABOOST REGRESSOR: \",adamax[0],'m')"
      ],
      "execution_count": 46,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "MEAN DISTANCE ERROR USING ADABOOST REGRESSOR:  12.979052804546736 m\n",
            "MAXIMUM DISTANCE ERROR USING ADABOOST REGRESSOR:  26.991121482947385 m\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "RLW_PO_7mdEl",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "b1cfe881-f7eb-4287-f233-44e0f1a3ed0b"
      },
      "source": [
        "count = len([i for i in total_dist_err if i  < 2])\n",
        "low_error = count / len(total_dist_err) * 100\n",
        "print(\"Percentage of data with less than 2m error : \",low_error,'%')"
      ],
      "execution_count": 47,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Percentage of data with less than 2m error :  15.636363636363637 %\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "SI4ibXg8roXy"
      },
      "source": [
        "### 2. RSSI Fingerprinting\n",
        "\n",
        "Fingerprinting of RSSI gives us a map of the RSSI throughout the area of inspection, it is effective for both indoor and outdoor experimental conditions. However, many data points uniformly spread throughout the area, gives the model a better understanding, which is not available in this scenario.\n",
        "\n",
        "### KNN - Regressor\n",
        "\n",
        "We have used KNN regressor to predict reference point locations using the concept of RSSI Fingerprinting."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ft_uMU2zdpiO",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "4b7f60f8-60cb-497e-9259-721a2a18c318"
      },
      "source": [
        "from sklearn import neighbors\n",
        "from sklearn.metrics import mean_squared_error \n",
        "from math import sqrt\n",
        "import matplotlib.pyplot as plt\n",
        "%matplotlib inline\n",
        "\n",
        "rmse_val = [] #to store rmse values for different k\n",
        "for K in range(13):\n",
        "    K = K+1\n",
        "    model = neighbors.KNeighborsRegressor(n_neighbors = K)\n",
        "\n",
        "    model.fit(newtrain_X, trainY_final[['Latitude_OP','Longitude_OP']])  #fit the model\n",
        "    pred=model.predict(newtest_X) #make prediction on test set\n",
        "    error = sqrt(mean_squared_error(testY_final[['Latitude_OP','Longitude_OP']],pred)) #calculate rmse\n",
        "    rmse_val.append(error) #store rmse values\n",
        "    print('RMSE value for k= ' , K , 'is:', error)"
      ],
      "execution_count": 48,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "RMSE value for k=  1 is: 0.00014264030700196525\n",
            "RMSE value for k=  2 is: 0.0001349883207070322\n",
            "RMSE value for k=  3 is: 0.00013347786765541687\n",
            "RMSE value for k=  4 is: 0.00013596747987660657\n",
            "RMSE value for k=  5 is: 0.0001352648427235548\n",
            "RMSE value for k=  6 is: 0.0001351045857036459\n",
            "RMSE value for k=  7 is: 0.0001342199386390155\n",
            "RMSE value for k=  8 is: 0.00013576804535908075\n",
            "RMSE value for k=  9 is: 0.0001366785072172612\n",
            "RMSE value for k=  10 is: 0.00013611592275921666\n",
            "RMSE value for k=  11 is: 0.00013672408536584632\n",
            "RMSE value for k=  12 is: 0.00013615789458791352\n",
            "RMSE value for k=  13 is: 0.00013672629676039693\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "2DH7ytXimluq",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 282
        },
        "outputId": "d880ee8a-a9df-4929-b63f-00d411cb33c4"
      },
      "source": [
        "#plotting the rmse values against k values\n",
        "curve = pd.DataFrame(rmse_val) #elbow curve \n",
        "curve.plot()"
      ],
      "execution_count": 49,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.axes._subplots.AxesSubplot at 0x7fa7457a4790>"
            ]
          },
          "metadata": {},
          "execution_count": 49
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "p11Da7-QnQHu",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 295
        },
        "outputId": "3e31550b-b2b4-4e4f-af62-0c5a618153f7"
      },
      "source": [
        "model = neighbors.KNeighborsRegressor(n_neighbors = 3)\n",
        "model.fit(newtrain_X, trainY_final[['Latitude_OP','Longitude_OP']])  #fit the model\n",
        "pred_knn=model.predict(newtest_X) #make prediction on test set\n",
        "\n",
        "x_ax = range(len(testY_final[['Latitude_OP']]))\n",
        "plt.plot(x_ax, testY_final[['Latitude_OP']], linewidth=1, label=\"original\")\n",
        "plt.plot(x_ax, pred_knn[:,0], linewidth=1.1, label=\"predicted\")\n",
        "plt.title(\"KNN 3- neighbours\")\n",
        "plt.xlabel('X-axis')\n",
        "plt.ylabel('Y-axis')\n",
        "plt.legend(loc='best',fancybox=True, shadow=True)\n",
        "plt.grid(True)\n",
        "plt.show() "
      ],
      "execution_count": 50,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "eqH9EPOGn2Q3",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 295
        },
        "outputId": "ead1de33-324b-4676-d9e2-0572e6256989"
      },
      "source": [
        "x_ax = range(len(testY_final[['Longitude_OP']]))\n",
        "plt.plot(x_ax, testY_final[['Longitude_OP']], linewidth=1, label=\"original\")\n",
        "plt.plot(x_ax, pred_knn[:,1], linewidth=1.1, label=\"predicted\")\n",
        "plt.title(\"KNN 3- neighbours\")\n",
        "plt.xlabel('X-axis')\n",
        "plt.ylabel('Y-axis')\n",
        "plt.legend(loc='best',fancybox=True, shadow=True)\n",
        "plt.grid(True)\n",
        "plt.show() "
      ],
      "execution_count": 51,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "VMdH2Mycn_kX",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 295
        },
        "outputId": "2ef55b2f-7736-4516-a2ad-d77bf91ea8f6"
      },
      "source": [
        "pred = pd.DataFrame(pred_knn, columns = ['Latitude','Longitude'])\n",
        "pred['dist_err'] = (((pred['Latitude']-testY_final['Latitude_OP'])*111000)**2 + (((pred['Longitude']-testY_final['Longitude_OP'])*111000)**2))**0.5\n",
        "pred['distance'] = (((pred['Latitude']-test_semi['Latitude_AP'])*111000)**2 + (((pred['Longitude']-test_semi['Longitude_AP'])*111000)**2))**0.5\n",
        "x_ax = range(len(test_semi['Distance']))\n",
        "plt.plot(x_ax, test_semi['Distance'], linewidth=1, label=\"original\")\n",
        "plt.plot(x_ax, pred['distance'], linewidth=1.1, label=\"predicted\")\n",
        "plt.title(\"Original vs Predicted distance for KNN Regressor\")\n",
        "plt.xlabel('Observation Number')\n",
        "plt.ylabel('Distance')\n",
        "plt.legend(loc='best',fancybox=True, shadow=True)\n",
        "plt.show() "
      ],
      "execution_count": 52,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAX4AAAEWCAYAAABhffzLAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nOydeZgcVbn/P291z2Qy2fcEQhIIWYCEJYQlhH2RHQUBFZXwEwX1uqIicK+KCveiosJVRFEQFBFZBAG5GAFZZQk7ISSEwEDIQhKyZzJLd53fH6d6ppdT3XW6q7p7ZurzPPNM9+nTp07X8tZb3/Oe94hSipiYmJiYvoNT6w7ExMTExFSX2PDHxMTE9DFiwx8TExPTx4gNf0xMTEwfIzb8MTExMX2M2PDHxMTE9DFiw18GInKJiPwu7LoB2lIismsYbVUbEZnk9T/pvf8/EZlXhe1eKiI3W9Tv2sci8msR+U50vYsWEfmCiLwvIltFZESt+xNTP/R5wy8i54jIqyLSKiKrReRaERla7DtKqf9WSn02SPs2dWuNty/SnqHYLCIvichJUWxLKXW8UuqmAH1qEZGjo+hDKZRSn1dK/bBUvVr20Q8RaQB+BnxIKTVQKfVBCG3m/E4R+biIbBCRw7Ju7PfnfedmEbnUe324V+dXeXWeEJFzfLZ5qYh0eufkRhH5t4jMqfS39HX6tOEXkW8APwK+BQwBDgQmAv8UkUaf7ySr18Oa8JRSaiAwFLgeuE1EhuVX6gP7oaczBmgCXrP9omiK2gbvae0a4ESl1KNZHx0gIgcV+eo24NMiMsmiS3/xzsmRwL+A2y2+G4ggvzlqRCRRrW31WcMvIoOB7wNfVko9oJTqVEq1AGcCk4BPefUuFZE7PM9lM3BOvnwgImeLyDsi8oGIfCfbM8qum+UVzRORd0VknYj8Z1Y7+4vIU55ns0pEful3A8r7LR8Tkefyyr4uIvd4r08QkUUiskVEVojIN0u1qZRygRuA/sBkn/0wRESu9/q6QkQuy5y8IpIQkSu93/gWcGJe/x4Rkc9mvf+ciLzu9XGRiMwSkT8CE4B7PY/vQq/ugZ7nt1FEXhaRw7Pa2VlEHvXa+SfaWBTbd9/y+r9SRD6T99mNInKZ93qkiNznbXO9iDwuIk6RPt4u+glyk4g8JiJ75LV7jYj83evnMyIyOevzPUTkn9523heRS7xyR0QuEpFl3rl2m4gMN/ymqcAS7+1GEXnYKz9IRBZ4fVqQbaC943G5iDwJtAK7FNln5wM/BY5VSv077+MfA5cX2eUbgRuB7xWpY0QplQL+BOwoIqO8vpQ6B3/qnYNvi8iXJFduLPjNIjI9a98vEZEzs3638TryOze8z3bztrNRRF4TkVOy2rtRtMJwv4hsA46w3Sdlo5Tqk3/AcUAKSBo+uwn4s/f6UqAT+Aj6RtnfK7vZ+3x3YCtwMNAIXOnVPzrr+5m6kwAF/NZrZy+gHdjN+3xf9FNH0qv7OvC1rH4pYFdDf5uBLcCUrLIFwMe916uAQ7zXw4BZPvvkHOAJ73US+KrX7hCf/XAX8BtgADAaeBY43/v+54HFwE7AcLSnpjL7G3gE+Kz3+gxgBbAfIMCuwETvs5bMvvTe7wh8AJzg9eMY7/0o7/On0BJHP+BQr/83FzkH3gdmeL/hlux9jDZQl3mv/wf4NdDg/R0CiKmPXtlngEFeP64CXsr67Eavz/t7+/lPwK3eZ4O84/UNtMc+CDjA++yrwNPAeK/d3+Cdp4bfNilvfw8HNgCf9rb5Ce/9iKzj8S6wh/d5g6HNFuBOb5/t5bO9Qd6xzJz/NwOXeq8PB94DxgKbgWle+RPAOT6/41K6r59G4ApgXdbvKnUOLvL21zDgQQrPwezfPARYDvw/7/0+3rZ2L3Yd+Z0b3us3gUu8vh+JPh+nZZ0Hm4C56HO5qVr2r896/GhPcJ3SXkQ+q8j1FJ9SSt2tlHKVUtvz6p4O3KuUekIp1QF8F31yFeP7SqntSqmXgZfRNwCUUs8rpZ5WSqWUfvr4DXBYqR+ilGoF/oa+mBGRKcB04B6vSiewu4gMVkptUEq9UKS5A0VkI7Daa+9UpdSm/P0ADEYb368ppbYppdYAPwc+7tU9E7hKKbVcKbUefXH48Vngx0qpBUrzplLqHZ+6nwLuV0rd7x2PfwLPASeIyAT0zeM7Sql2pdRjwL1Ftnsm8Hul1EKl1Da0kfGjExiHviF1KqUeV97Va0IpdYNSaotSqt1rdy8RGZJV5S6l1LOq24vd2ys/CVitlPqpUqrNa+MZ77PPA/+plHovq93TJZjsdiKwVCn1R+/8+jP6xnxyVp0blVKveZ93+rRzDPrm86rP59vRHv9lfh1RSq1GG8ofBOg3wJneObkd+BxwulIqJSJjKH0OXu3trw3om0Y+Xb8Z7Qi0KKV+7+2DF9E3ujO8un7Xkd+5cSAwELhCKdWhlHoYuA/vOvX4m1LqSe9cbgu4PyqmLxv+dcBIn4tmnPd5huVF2tkh+3PPCJcaSFud9boVfXIgIlO9R8bVouWU/6aEVJHFLXSfUGcBd3t9Afgo+gJ5x5NBig2OPa2UGqqUGqmUOlAp9WDWZ9n7YSLao1nlPcZuRN+oRnuf75BX38+Qg34qWFbsx+Vt94zMNr3tHow+ZjsAGzwjHmS7Nn38Cdp7my8ib4nIRX4VPYnhCk+S2Yz2lCH3WBrPAYrvi4nAXVm/+3UgjdbzS7EDhb/vHfQTVIZi53mGLwBTgd+JiPjU+R0wRkRO9vkc9NjasSKyV4Bt3qaUGor+nQvRT8Zgfw6afl/+OX1A3rn1SfQTCvhfR37nxg7Acs9RylDOPg+dvmz4n0LLLKdlF4rIQOB44KGs4mIe/Cr0o2Tm+/2BckPnrkV7YVOUUoPRj4h+F1c+/wRGicje6BvALZkPPE/6w+gL4m7gtjL7l70flqP330jvRjFUKTVYKZXRslehjViGCUXaXQ5M9vksf98vB/6Ytc2hSqkBSqkrvG0OE5EBAbcbuI+e5/0NpdQuwCnABSJylE8fzwI+DByNlg8meeVBjuVy/PX15cDxeb+9SSm1IkC7K9GGLZsJaFkmQ5BUve8DR6HljF+ZKnhPvt8HfojPb1Y6yugqr04glFLrgPOAS0VkHMHOwfFZTexEIfnn9KN5+3egUuoL3vaN11GRc2MlsJPkDhqXs89Dp88afk+++D7wCxE5TkQaREca3IbWIf8YsKk7gJO9gbNG9ON3UGOdzyC09rlVRKajvatAeI/mt6O9j+HoGwEi0iginxSRIV6dzYDr31Lg7a0C5gM/FZHB3sDjZBHJSFO3AV8RkfGio4J8PWS0h/hNEdlXNLuKSMZIvU+uIbwZvb+P9TzrJtFhguM9eeg54Pve7z6YXCkjn9vQg9S7i0gzRQYcReQkr1+C1mXTdO/H/D4OQhukD9DjL/9dpA/53AeME5GviUg/ERkkIgd4n/0auDyzb0RklIh8OGC79wNTReQsEUmKyMfQ41P3WfQNAKXUSrTxP05Efu5T7Y/oMYrjijT1M+AgYDeLbS8B/gFcGPAc/KqI7Cg6RPvbJZq/D72PPu3ZgwYR2c8boPW9joqcG8+gn+Yu9No6HH0+3hr090ZFnzX8AEqpH6O96ivRB/IZ9F3/KE9DDdLGa8CX0QdzFXqgdw36wrflm2hvcQt6APgvlt+/Be1l3q5yxy4+DbR4ssPn0Y+vYXA2etBqEXqg8A605AK6//9Aj2G8APzVrxGl1O1oXfgW9G+/G33zAj028F/eo/c3lVLL0d70JcBa9PH6Ft3n8lnAAcB6tCH/Q5Ht/h/a63wY/aj+cJHfOgU9OLgV/bT4K6XUv0x99Lb5DtqzW4TWxAOhlNqC1tFPRstBS+mO9rgaPW4zX0S2eO0eYGrH0O4H6PGDb6BvSBcCJ3letDVKqXfRg5Wni0jB+I1SKo0e7yqIOsqqsxkdBeRbx4efAOeJyGhKn4PzgVeAF9E3vxTaMJv6swX4EHqMYCV6//8IPZAO/teR8dzwnnxORisI69BPSGcrpRZb/t7QyUQlxISEJxVtRMs1b9e6PzExMRoROR74tVIqX/Lqc/Rpjz8sRORkEWn2tOUr0REPLbXtVUxM30ZE+ouOvU+KyI7oJ8C7at2veiA2/OHwYfSj4Ur0Y9/Hi4X6xcTEVAVBj+NtQEs9r6Plpz5PLPXExMTE9DFijz8mJiamj9EjEm2NHDlSTZo0qdbdiImJielRPP/88+uUUqPyy3uE4Z80aRLPPfdc6YoxMTExMV2IiHE2eqSGX0Ra0HHZaSCllJotOpvgX9CzGVuAM708GjExMTExVaAaGv8RSqm9lVKzvfcXAQ8ppaag0yIUm9EZExMTExMytRjc/TA67THe/4/UoA8xMTExfZaoDb9CTy9/XkTO88rGeDk2QE+JNmYWFJHzROQ5EXlu7dq1EXczJiYmpu8Q9eDuwUqpFV5OjX+KSE6OCqWUEhHjRAKl1HXAdQCzZ8+OJxvExMTEhESkHn8mXay3QMJd6BWH3vdSquL9XxNlH2JiYmJiconM8IvIABEZlHmNznq3EJ1dcJ5XbR565aiYmJiYmCoRpdQzBr1aUGY7tyilHhCRBcBtInIuOnXtmUXaiKmQhSs2sXl74Sp608YOYsTAfoZvxMTE9HYiM/xKqbfw1pLNK/8AvYhDTBX4yDVPMnvSMCRrbZiVm7Zz3IyxXHx84PUvYmJiehE9YuZuTPmkleKWzx6I43Qb/useW8baLeWsExMTE9MbiJO09XL8kq/GSVljYvouseHvA0jeCsBS9pLAMTExvYHY8PdRYoc/JqbvEhv+PoDkufz5TwAxMTF9i9jw92KKra4Wa/wxMX2X2PDHxJTL746Gv55f617ExFgTG/5eTDGvXsUqf+W8twBeubXWvYiJsSY2/L0ck56fr/nHlMHWOMVUTM8lNvy9mGI+fazxV8iVU2rdg5iYsokNfy8n9u1jYmLyiQ1/L8Yvqie+GcTE9G1iwx8TExPTx4gNfy9GYR7IFSke4x8TE9O7iQ1/TExMTB8jTstcYza2dnDtI8tIu7keuAice/AujB3SVHbbSpn1fCHO1RMTUwu2tHXy6opNBeVNDQlmTRhWtX7Ehr/GvL5qCw+8tppPHTAxp/wvzy1n/51HVGT4Y2Ji6os7nn+P3zz6FpNGNueUP//OBp66+ChGVmlVvNjw15i2VJqJIwbwuUN3ySl/5u0PKtbhFcp3Alcs8cfEVJ+0qzhh5ji+e/LuOeX7Xf5gwVN/lMQaf41p73RpSsaHISYmpnrEFqfGtKfS9GtIGD6RUHR406IrInGunvCIZ0XE2GF8Cqe6s+ljw19j/Dx+HXJZWduxnBMTU1/4XZPVdsZiw19j2lJp+jVEeBjqwLuIiYnpph6eEePB3RqjPf5CqUefHLF1jomJiieWruMPT7UUlA/ol+THp+9JQyJ8h8zPqxeqG3ARG/4q8vqqzbR2pHLK3lq3jeEDGgrqhiH1gI93IeGMH8TE9GQWtKynIelw8p475JRf9vdFPPT6GnYa3j+nfPiARsYNyS0rh3rIih4b/irR1pnmpF88wV7jhxR8dv5hkyPZZiznVIN4J/dUlFJMHT2I42aMzSl/fdVmrn5oaU5ZRyqNiPDgBYdVuE1zudb4q0ds+KtE2lX0Szr89YtzA9WXsKJ6Yo0/JsaIq6DBcH18/ZipfP2YqTllb6/bxjm/f7bibfrmz6K6+bPiwd0qobAb1Aklqif2RmNifHGVwnHqQHepAbHhrxJKqZoseegXxx8T09dJK/PMdhNhPSX75s+q8mz62PBXibI8/go99uInUvw0ENO3UQoSAS1/b5v0GBv+KqEsLb/JUy8Hs8Yfu/wxMa6rcCwef0Px+PFx+atMbPirRRnHu3KNP7q2Y2J6Oq4KLntG7SyFFb4dlNjwVwmdKdPK5Q8pV4+hrA48jpiYWuOq4B5/WIZZa/y1H3eLDX+VUBbeBYTzNFgsPCz2+GP6Okop6imop5pjCJHH8YtIAngOWKGUOklEdgZuBUYAzwOfVkp1RN2PWmM7uAvhxPX6xQzHhMd//OmFgrLTZ4/niGmjK2r33Q9a+cRvnyblujnlgvDLs/Zh9qThFbXf13EVNQnn9Bt3620pG74KvA4M9t7/CPi5UupWEfk1cC5wbRX6UVNswznDCP0sqvH3ogiFWnP8zNyZn/9avJbH31hXseFft62doc0NXD9vv5zy/7zrVVZvbquo7ZhMOKeN1NN7rplIDb+IjAdOBC4HLhC9l48EzvKq3ARcSl8w/NTG0441/ug5KS/Xy8bWTl5bubnidtOuoqkhUbD8Zv/GBFVcrKnXopSyCOcMZya9Usr3mqzmIY1a478KuBDIPKuOADYqpTKZyt4Ddoy4D3VBORp/lPn4e5HzUnc0NSRo70xX3E7aVSQMUoSe7BMfwEpxXaw0/ih3ea9J2SAiJwFrlFLPl/n980TkORF5bu3atSH3rvooS58/NK88juOvOv0bErSlwjH8SYNlcqoc+tdbsYrqCWmbtg5gVETp8c8FThGRFvRg7pHA1cBQEclITOOBFaYvK6WuU0rNVkrNHjVqVITdrJDFf4fOAHprGQe8Yh2+mMdfWcsxRWhqcGjrdEtXLEHKx+N3RHBjy18xVnH8Ec/cDUtKCkpkGr9S6mLgYgARORz4plLqkyJyO3A6+mYwD/hbVH2oCreeBcf/GA44v2g165QNRJmP31y3rTPNy8s3FpQnEw777DS0zya0sqWpIcH2jso9ftdX6iHW+ENAWXn84UTdKHzi+KnuU1wt0jJ/G7hVRC4DXgSur0EfwsVNlaxie1DDieqxi+N/YOFqfnjfIiaPGphT/uqKTdzzpblMGTOo4j71BZoaEmxtT7FmS+GT4ND+jTQa1lg2kfKReoTY4w+DtFI4fXQmU1UMv1LqEeAR7/VbwP7V2G7VCHAR6pm7oTdbEps4/s60y2HTRvGzM/fOKT/h6sdpT1UuXfQVxgzux7qt7Zxw9RM55e2daU7ZewcuP3VmoHbSrusj9RBrdSHgKuxm7oawTV+NX6CaBzVeiCUUAhh+n6nafoRxbReN6rFovSHp0JmODX9Qxg9r5qmLjyoov+vF93hkSfBAhbRLrPFHiO3gbtRRPdWkjz7ohEwgj99ycDekM8E4S9CnI376Y4MjdKZjQ1MpScchZbEfU65LwqBFOE6s8YeBjcbvfaPybWKO49f9qbj5wMSGPxSCePz+B7zYdyLD1LTPY2hDwiEVe/xGXBX8qDYkxOrJyS+cEySeeR0CVnH8Ebvk1Y7qiQ1/GATx+JXdgG0Ya+76RRLZnsMNSYeO2PBXTNJxSFm46imffPFOHNUTCq5NyoawonqsH/2jIdb4QyH8qzDqc8Ps8JufShocsZIoYswkLT1+13cCV9+ZufvHp99h8/bOgvIpowfyoT3GGr4RHD24G6xumCkV/JyxWOrpaQT2+G3bLa873du0Swzn18eGRDy4GwZaMrPz+BMJnzj+PuDyp13Fd+5eyLb2VM5fy7ptXP3Q0orbtxncDQu/o1btpR1jjz8Uwg/njPJ0tM00mEwInX3A0JSFxYFKOlKQYrkYadecRMypsh5cK5TSE9guPG56TvnCFZu48I5XKm7ftYjjDzOXTh0oPbHHXy1swzkhhMXWsbuB+EX1NCYcOuM4/orRUk/wY+qfpK1vaPx+529Y4ax2cfwh3Wx9+l3tfPyx4Q+DCMI5o1yD00+v9JN6bLXpGDN6cLfyqJ6+pPGbcJywlkGs/mLr4L/0Ym9P2dALCT+cM4wMmmGdSA0JJ5Z6QiCZsBsk90vSVu2BwFrh54iITzirUopH31hrnGV+wM7DGdrcmFNWk+ycIbVTKbHhD4MAR1N7/LZST+XYLPPmNw7RkHD41+I1BdEVjghnHTCBIf0bQuhp78dvkNx1Fd++8xU2t+Xu37fWbuNDe4wpqO84fWPmro4yCx7OumpTG+f/8XkOnZqbzXfp+1v45AET+dyhu+SU28Txh7kCl58ZiAd3exxBUzYEJ4xHP9sTSW+vsJen7L0DDy56n23tucno/vbSSvYaP4SDdh1ZQS/7Dnpwt/CYrG/t4IGFq/nJGXsWfLbPhGEFZT1a4/9gGbSuD1RV0mn2kqWwPPf8atrQyrTUYljenFOe2NzGYc0tXHdEbv07XtjIttXrYPkHOeWT2xcxaN1W6DekZF+S7Slm8gYsr2yd47Gb36Eh4UBeFtzpqcX0fz8N6dwEiYgD4/etaJsmYsNfCRnLHMhC21n+8Eb+ffREixZmTRjGLIMBevHdjXXz6NoT8AvnXLulnXFDmzhuxrhA7fTo7JzXHQ4N/cEp/ZTYgOLqRBvc/puc8nGuy3e3d8DtuUtSjnBdfthZWH58R4qOlIt6L3ebX93aweDHkpBMlOxLs1JcSRvc3r9k3WKc3Napr+23c/tyybZ2hj7YAIm8YdeGJvhyWWtZFSU2/JWgMo/t4Xv8utXoFmLxq25zwwlrkK2v0NSQ4P3NbRx/9eM55a0dKSaOGBC4nR69LEKqHc5/FIbvUrJqe2eaI74/nyUXHJ9TvnzdNv7f75/lkQuOyCl/b9025v3+WR7NK3/1rQ/49PXPoFpz23cc4Z7z5jJ97OCSfdmyvZMPXfEwr15wbMm6xfjNPxbTlEzw5aOm5JSfc/Xj/Pj0PZmxY+mnjzCIDX8lZAx/4Kgeu+HdcNIy+/THKPLbD0DHOWOCM2pQPx742qG0G5ZlHDu4yfANM45ID57ApbR8UQF+Gr/fYO2Bu4xg6eUnVLTNUGfu1sGNOzb8lRChxx/GyRG1aah2CFpvYNfRA0tXKkGPztWjXIJeCX5RPX5x/DZLKZZDGIO7fk1U+1qK4/grwcrjL2MhljK6lI8xL4jPZJRy8kf1WK25JyM9+ElLBff4/aJ6Ms0UlkWXgiHMVo2LI1U5ZUNs+CvB2uO3W4ilUmxtsnUf+0jqgHqjx3v8QQ2/z290HPMENpuka7aEda7Xy2GLDX8l2Hj8lo+hEtIsHXMcP75noNXgbhjLhMVY07Nn7tpdCGapx07jD4PoV+CKUzb0HGw8/jIsZKXngX0cv119KWMbMZUj9FCJLdPnwFKPXa4em/z6taKoxl/FfsSGvxIyhj/AhJRyFmIJA9+8IKYp79gOQAsWqWdiQkJLHbXuRRl0OUpBB3fNhlzf+Ez1o5R6wnFyyhnri4LY8FdC5up786FA1a3j+CsN4y/j+3Y3p1jpcWqwB3rszF1Lj98PfY6aPf7opJ5ob7Zhpn0OQmz4KyHjwST7la5ahsYfxongl6vHhO3mpEdrzT2XHqvxZ66XiqUeP40/2sltoexxywCKqIgNfyVkTuRUW+mqNViIpdiJak7SZke1dckYTY/V+DNnS8ALQflY/lpo/JHLM/Fi6z2IjOHv3F66ahl3+uji+IvUt7w59UjPs4ejPf5a96IMujz+gCdZkYFQ08xlHcdfZt8q6I9tE36RdnFUT0+hy/C3Fq+H/eQoCeHiLmaU/SbA2BDP3K0NPVfjtxvc9avpF1Nvs6KWLdWeYBU1ccqGSsg2/K5LsQU8bRdiCQu/qAjf+ha97Ctrv9YbYS09aKIz7fKXBcuN6wacOHMcoy1yChVgHc5plm4cH4fDdau/eLotfnZAik2uiYDY8FeCyro4Utuh0T/DogJrobDiOP4iDfh5L/WyPGRPwUWq/tgcpW1754NWfvzAYk6bNT6n/KllH9CQcPjUgRPLb9x2cNdv5m4NcvVUJ6onuvbziQ1/JWQb/nQHUMTw1yBJm23b1lE9PTkvfA8mSo9fKcXowU1cesoeOeWX3beI7R2FWUUtW9f/Kpy56+dwRJqrJ6RABtvovqiINf5KUG73ghLpVPG61lE90Ybs+S69aNNIHNVTE3Q4YzR73i8ksrlfkm0dpc7xEoQWztlzs3P6Ue28V7HhrwSlIOEZ/rcfLVm1Nhq/sdSyvk9d4qiemiAS2eBu2kcnH9CYoLVSj7/rXAk+c9eEr8dP/Wfn1Dczu3G3KIilnkrITjF757kw83T/qljOio34TDBdUrY2vN4H0gDWbG7jigcWF4T/iQhfP3oqE0Y0+3yzfvEb3AwDv9mvzf2SvLu+dPRaUcqYuWse3K2Bxl8FjzzW+HsKKp17EhcR8Kw1fsJJ2eCnkRrrY39zqneNf9nabby8fCNfOnLXnPK/vrCCe19Zyan77JhT3ph0GDmw9EzsWpJ0hEeWrOH//f7ZnPKGhMNPz9yLQU2l17P1w1XKGJw2oDHBgpb1/M/9r+eUJxzhC4dPDrbNkKQev1QhPSE7Z7FrsppPz7Hhr4R0Z7fUAzqs0yeyRyecsms+yrhh3yyBFm1UOxKhHJRSjB7UxKn75EapNCQcLv/769z89Ds55Wu2tPPsJUcxoo6N/wkzxzFqUGH/LrrzVdZt7ajQ8Juf5A6eMpK1W9oLzsgbnnibE2aOC7hWrP3MXZsVuCKfwNWLiMzwi0gT8BjQz9vOHUqp74nIzsCtwAjgeeDTSqmOqPoRKekOSDR2v9++wd/wYxcjH87Si+YVjPyaLi9Xj32/qomrzNMrTtpzB07ac4eC8gP++0E6DDHs9cSgpgaOnD6moLy5cVHFbft5zaMHNXH+YZMLyu99eWXwc8Da4y+u8edn73TdaCdwhYHvb6L3DO62A0cqpfYC9gaOE5EDgR8BP1dK7QpsAM6NsA/Rku7I9fi3b/Ct6pd3pBjRGlWDx4SdHtUTcvXYPv5Xe0GMMAkjaZ6eBBVSh/KxTdmgKxeWiBgHeKPN1aPbjUyOqfKcmMgMv9Js9d42eH8KOBK4wyu/CfhIVH2IHJPH74NtqGQYg0n+emKRqB6r6fT1H8dvawx6wriFH37atw22aQ+sUhnYej9FBmtNvzXq7JwQ1rhb7fWowFKPiEwEpiilHhSR/kBSKbWlxHcSaDlnV+AaYBmwUSmVCQh+D9jR57vnAecBTJgwIWg3q4uF4bdccS7y8C7zBBi7NnzSotcVtotz2I5bbJWBDLbuVUSE4DXqwd3Kn5C++7eF/N/C1TllY9Ra7rGY6Vzsp2R0/kTWlRLlBK6o0ZdSnQ3uisjn0EZ4ODAZGA/8Gh1Mff0AACAASURBVDiq2PeUUmlgbxEZCtwFTA/aMaXUdcB1ALNnz65P85I/uPvKbbD2Df26/1DY77Nd1t5W44cKL+K3H2PI4sc5u+M9eG8wjJgMz90Abppd1m7lpI0b4dFHuus3DwMOt4/jr3PLby31WBqOVYkd6sfwA5XeicOSet5et43/OnE35uwyoqvsnWVLcP9ml+LCryumAV6/8ZywCEva9M3VU4fhnP8B7A88A6CUWioio4NuRCm1UUT+BcwBhopI0vP6xwMrLPtcP+R7/E1DYUMLpNvh1dthn09BQ3+gjKnalV58T11D/83rOCS9Hp4TmPFRePIqmH4y/bduZ2RqK2zwHry8/jqHPGWVeaYn5Oqxffwv+ps2rYCl86Fja3f9OrrxhRFlVZ7UY2Zoc2NOUrdN/RO4Fid20Wumyhq/t0lP4++ZTxXZBDX87UqpjsxOFZEkJe5PIjIK6PSMfn/gGPTA7r+A09GRPfOAv5XZ99qT7sw1/B+5Rv/fvlEb/qxcPvYLsUhl3rSbZstuH+Ox9S8zVbm6L0N2go9cw+uvv8+fnnmXQz6yn9ffDbq/bhqbYR8nwhmkYRGqxv/yn+Hpa6F1HQDbRu6Fs3FbGN0MhTDGhVylSFhJPeYBT9O8FcHFxmgWO/9Nk9iijOMPC7+w7mpH9QQ1/I+KyCVAfxE5BvgicG+J74wDbvJ0fge4TSl1n4gsAm4VkcuAF4Hry+x77cmP6sngJPT/bMOvqiz1KBdwtIel0vp91hmXc7Fmwuss5xrY5ihfuXE7C1dsKigfNqCR/SYND75hC2xju4tegErBjvvC0n/AAZ9n5cD9SDx6eRjdDIUwTF460sgYZeXxg/81Y5J6bMdzbAlr9q6v1FNFghr+i9Bhl68C5wP3A78r9gWl1CvAPobyt9CyUc8nX+rJ0GVIsz1+y8HdSk8E5aJEUDi6H8r1j5/2ykW5ljcnu9DHax9ZxoKW9Ywf1p0mwVWKfy9bx+IfHm+x3eDYShdO0RmUKmcfKnFqsti6H2FIb0opEnbhZ+b0H4YnXFFYxbYVk3rMGn+0Hn8os+mLfVaHGn9/4Aal1G+hK1qnP1Bh8o4eTkbqcZLgZmUuNBl+29WtKu2bclFOAtdg+At02S5jZtlHy8Gube0pzj14Z86YvVNXWXsqzczvzbfarg1+Scf8KOrVKdX9NAe4OHWm8VcoD2I/CarYZMB8J0L3LriUWOyXCIWrkEWdnRPCCWYwLo5k+fRcKUEN/0PA0UBmVKs/MB84KIpO9RgyUs9/PJu7/GKWdJJBe/y2Uk8FJ4JyUTgokcAeP8q1yZ9VwjsupLUjTXNj7ikXhrEqhmsrX1HM88q1LArBoX5m+Ybh8Zcj9QTdpuBaDu76Px+YomAi9/hDaLpegiGCGv6mrMlYKKW2ikjPS2sYNhmpZ0TeVPaM9XSz0tjaTYoNQerRskSux6+91YLY60y5slNgbWe5tnamaW5M5JTp3PIWG7VEKawGKyl2M8vOxoqWeqRermSPUKQei5u/Pk99BncLpB77iBi/m1Ay4XDeH5+jMdnd2dWb2th34jCr9q36EtKsbt/B3TqUeraJyCyl1AsAIrIvsD26bvUQ8uP4M4hhcLeshVgq6JvSERQKC4/f0nu1zSi4vSNlMPzRzv61T9lQTGJQ3ccWUDh15fFD5XJBOiypx6Dm23v8/p/94TP7s35bYYqvqWMGBW7fmjA8/mL5hypvPjBBDf/XgNtFZCX6548FPhZZr3oKvoO73hlSENVjR0UngnJR4niGX+Uafj+N33UhGZaRLMQo9UQ8F8BW93WKJZ7L9/jrTuqp3GssRy4JnqTNeo03X4JlA61P6iHgNJDhV0otEJHpwDSvaIlSqjO6bvUQ0h2QMKTvFYGMp+1hm6MjjKgeEJRkpB5VxOP35meUkU/IJNM8smQNF9z2ckH5pu2dDG3OfULKTn4VRRih/czdYl5z7j50pd4GdyvHOmWDX1SPQdURsTf89RSWH0pUT5Hv12s+/v2ASd53ZnmZAP8QSa96Cn5SD2gDkR/Oadl8pVKPKtD4MwY+P46/8EYVBD+pZ9WmNg6dMpLvnLR7Tnky4TCkf+H+Ek/ntwojDEhZcfzFPH4nX+qpI8Mfwv5zQ8ppr8/3vKgeZRfHbz3bPWJCi7wxRvXU4QQuEfkjOkfPS0BmxFIBfdvwp9qhv88Ytzh64pSH7UIsoYRzdkk9JTR+6LpR2Y5DmEi5iuZ+ycCLmThd6YSj8PjtUxD4jznkD+6KNxu1PghDNisnnDN4VI/SUWYBqcc8UFHG8VeToB7/bGB3Fa+snYufxg/aM8zz+K0Mm1QY5ugZepPhN0oq4lhP4PIzNK6rSFq4jVFG9pSTvyWoxu/WmccPlRvLsqQxY6pXc1SPtdRTF4q4Jqy+mFrxCY6KjKCGfyF6QHdVhH3peZSUerLi+Mt4bK1Y6sEJNnMXvM9sPX6zd5yynTRFdJE9tkna/Pp98V9fYc/F7zBYbeFE4K4XV3Avb/GzevL4QwgJDE/qMUf1hDVztxaEEXlTL65zUMM/ElgkIs+iV9YCQCl1SiS9qhH/9+oqbntueUF5whEuPWWPnFQDgK/Hf9WDb/D5lOJH97zKmkadm2bVpu3GdVL9qFzq0VdN1wQut3theOPjed6NKgiOY9YlbT3+KC9u2xztfk8xL767kc9OGs6IZH9YBIdPG8WsaTMYfH+isHKNCMMwuZbzHuySi1lq/IFrVoeCsbFy2zE9cIc1fhCQoIb/0ig7US8827KesUOaOGb33PVMf/zAEt5d3xrY8N/5wnt8MZHgsCkj2Dp4bFf5NIsY48yAZ9motGFwt7jH76i05QQus5FMuYqExUhtlLH8tvnl/S5ApWBoc5KhDTrN8LDmRoaNHITt3IcoCcMwWUtjPlVNUWy2E7iKzdztufjE8VPdp4Gg4ZyPRt2RmvDyrbD8GZh0MMz4KErBlNGDChay/u1jb5uP1ztPwq6Fa9Gk04pkIsHhU0bAmMIFvQNTccqGTJK23Dh+o3ErQ+rBx0imXZdECF52GLi2YbSYQ1Rdpfdmzs0zL3Kr5oTw6FTOQiw+En+hxi8KV9mGc9aP6Q8vO2ftf1PQqJ4DgV8AuwGNQALYppSqr8WH8nn617Dgd3D8j4wGmgXXw/b18MEyvVAJ5mvHcQzet+vqPPZj9yyo3+kqSCQqMgpiPT0qD+WCT5I28wYd/Zss+2i66NMuloO7UWr84QxWukppjzU7D1OdGX7bM+bFdzewsTV3Os7i1Vus8/EPXLMAnv9zTvk3Nn7A5EcGQXP3E/GQzWv5wCIZVH1KPZW14ff9MCbf2RBU6vkl8HHgdnSEz9nA1Kg6FRpTj4XX74EP3jQb/nQHjNsbtugxa79HS+Pgo+tdMMN2Lqif0pavMsNfqV5bkKSte6TMaLBFq7U2HpZfkra061pOAoouqkdZDu76TkjqqpCl6ZcxLhIltk9OZ/32GWZPGlZg6E/fd7zVNoesfho2LYeZZ3SVv/7uW4wfPZbhw7vl0bbWDn6++jCuDNh2ObPdIycEw2/O1VPdG13gCVxKqTdFJOGto/t7EXkRuDi6roXA8J1hwKjcZGnZuCloHKBvAGQeT02TKwwHJe0ZfkNUTyqz+GctvcEcqSd4HL8NfoYmrRSNFoufRhnKZr94uI90ofCkHq8tEe8pyefcqhnBd2RaKX579myaGioboBY3DWP2gAO/0FV23wtPsvceuzFhYvcCO9s2bufJZ/5t2XhFXQuXeupLhQQ1/K0i0gi8JCI/Rod1RriscYgUM2jpjlzD73PNGKUI7zumwd1U2jMQFUk94czcNcfxmzZo7736RXSkXDtj6zjRhnNajVX63IVc09Nghcc4bKzPmTAmoSJ6oqJT2pTYn9P18zQFGa+8sj75zWSwTXhYKUGN96e9ul8CtgE7AadF1alQcYpo7V2Gv1vn9A21KjD8xTx+F6lQ/604vCtn6cVCj98Uzim2cfw+s1zTadsJXHWk8WOWnXS0S95dpNi5VQNsBx9tM8aatwmi0rkSGJnzSwrq2pzT9Sb1RKvDVzdlQ1DD/xGlVJtSarNS6vtKqQuAk6LsWGgU9fg7cwy/393YmLEx3aFPdqfwMVlLPTU2CkpPj1eS0HJEtsdvqp9J2WA1c9dncDfUVMiVYR/Oaf5Nrpu5NLMaq8fBXYsdGZoRc1MFHr8pqqcc6imqJwzqZVJaUMM/z1B2Toj9iI68nDk5pDugIU/qMRwVY0oBn4XW017Fyj3+EPLxF8nVYwrnFFuNH7PBTltP4IpW6rEJLS295m4dG37bMEwqDy0UwZN6Ch2g/JZtZxbbRf1HT5Q586NOT55PUWFORD4BnAXsLCL3ZH00GFgfZcdCo6TG35zl8fudaFJoDNyUWd93XRocJxSjUGlUj0tm5m5uHL9ZZHS8qJ7gm9AStymqR9mFBEYax28XqeQ3btGt8ecb/voZ3BXTeVrqO2F45SbDb+iHreGso4ApIPNEVanG3zPi+P+NHsgdCfw0q3wL8EpUnQqVYoOWaS+qxwvN9B/cDe7xp9KKZEJq7w12efzmqB5zygb7pReNGr+rSFhE9TgRGn69lKDNWoLmcYvuCVx5hl9vpD6e37E1rJXPjBVER/UYpZ48jR9LKSqEMYiwiczjj7T1QooafqXUO8A7InI0sF0p5YrIVGA68Go1OlgxJT3+gd0ROj4nmnlw15yuIZX2vF2pcAJXpcYwY/i74vhdHWKKj8dRRvip32Bdyjo7Z+VSz9vrtvGLh5cWXDuvrdzMSXuOC9yOX2ipq0DyByyzl9iUOsjZY3nO+IUvW20yM7hriOopjILKbNWi/TrwjjOEMd7gl7WirqSeLB4DDhGRYcB8YAF66cVPRtWx0PAz/EppT7+hOTeqx3BUHFO0hE9mzpTr0pBwQonxrjSqp1h2zoKWy4jq8fPUXctwTh1JU9lZ/9rKTSxbs5Wz50zKKZ+760gO2nVE4HaMx5qMd+zn8bvoyey1xTbcMLSoGTflE9WTi7XGX5dST617EQ5BDb8opVpF5FzgV0qpH4vIS1F2LDT8DHDG2DcO0AuqrH+b/p2bEBlaUNU3jt+o8XvebqWDu5Vekm46b+Zu6Th+ce2jelxV+DRk6/GHESanFIwf3sxHLWadmvtivrh1WX5Uj/e6TgZ4pYzwqKg0fpNMU47GX09STxgZNH3j+KsczhnY8IvIHLSHf65XVnsXJwh+Bjgj7wwYBU2D4Zf78XVnEH/f+VFDGwZjkO6kzXW485l3coo3tnZmGf7yD2U5F3EOXi4Z31w9Jo3fMtPksOZGfvKPJfz60WW5TQlWBtgJIfNBWBEgfhd31+Cur8dfe2yMR+ZmXbnUY9b4M/2JCUa9Sj1fQ6dnuEsp9ZqI7AL8K7puhYhfPH3G8DePgG+3wPq36PeLA81NGD3+Tta0Ku5/dRUThg/I+ej/zd0ZXqv94K4Sc8oGvzh+UXaDaWcdMIGzDphQcVe1warQk7JNJ1ykL+YJXJjj+AEe+R991Y7ZQ7/vbIVRu8HEORX3pyipDp2LKtkEnds5ZPsSRr31KmwdCKk2GLcn7Liv8auhGhmTx2+UeuyiYupv6cUQ4jlVOGMFlWKTlvnRrPdvAV+JqlOh4ufxuyn9P6PTOw0kVMo/gZIhqidFkk8dMJHjZxoGD1+vVOqpPJxTFYvqCSGOPyzCWHrRNhmbH8Wyc5Kv8TcOhIO/DptXwSu3wpCdYOsaSLfD0InwtYgD31a+CHee29WXQ2VXRi/rD/9+SpftMAvOi9Y/E8yDuyaZxnpmcZ0YyQyRx/FX8UZXNM5NRK7y/t8rIvfk/1WnixVSTOrJnnmbaCRBGsdQ1zEdlHQHKUn6n5gVxnhXnLsjJztnbhy/35q7tVpUJIyoHvvVXM34Zuc0efyOA0dfCsf8QL+f8VGYfKR+ne4kctxO7e0DjJzK/4z6Ea8e/UcYMcX7POX7VZ+5itaI4Cv1FNTF/kmjfsy+JipJstqyWKmj9Ufvf9BMqvWHiM/gbt7grOf5iyq8WByRwlT1booUSX8vs5Zx/OkUpLaDiDaHndtgQwsM3rGriukJZtT2t1leCw8rBH3Tdf3Xy7Xsik84p0Hjz5Atc1Rz/7lpSPTTso443VEzAQadw1zdSlQqd/wIs2dc7URkYRP1ka3mrinq8Sulnvf+PwosAhYppR7N/FWjgxXjF0+f7swz/Pp1wjV4aqZHPE/q8Z2hGkJUT9knwnvPAtDZbxjrnZEweAdYuwTGzNBtm7o8ZDwHr/5DTTwsnQup8hmRYXTe75Hb6PFnyBj+7B1bjRm9Kg3J7nO4QIooEk4cRgw/ZG6UJqnHENVjGbnSK6N6/MbRIpSRTJR8PhORS9FZOR39VlLAL5RSP4i4b+FQTOpJZP18z/A7qtDw+4VzpiTp72WGkp2zTFLtMHIabuMgPkiMhM8/UVCloO1Dv0Xb8nML6lUDHcdfWRvaIw/HkJkWIjNq/BkyRi/7HEm1Q9smaBpScZ98adtUEFKccwNNtUHndmjoX/DVsGL4u6N6TLl6Ciy/1UldbzN3K3LG6oxSuXouAOYC+yml3vbKdgGuFZGvK6V+XoU+VkZRw18o9SSMUo/hETXdqaUeX4+/htk5M7N2fVwmc1RPQidxrsGFpidNVS6ehjO4a6Hxd33JENncthGumACnXgfjZ3tRZKINZBHt3UiisTsKTbzJga/fCw99v3jfN7wNl4+FLz3vZdBMdOeY6tjOzrIK1rxu15c8duhoIdm5NZjGX8bx6W0hoX5jK7YRT5VS6mh9GjhGKbUuU6CUektEPoWewdszDL9R40/lzrwV0dKNQeox3unTHXSW0vgt17DNp+zzQCmjB5ZbpTCqhxB1XxtEKt5VoXmHftk5uzV+05e8yyi/A4N2gL9/Azq26EHYVJsubyqcJOiLcqF9s/5+sknfUHzw/fm/zArpbBwEHVtoBP7ZAFw/pCI95eKONA3JRM74EfjLNLZSTz0RbVRPfQ3uNmQb/QxKqbUiUpivIAsR2Qn4AzAGvb+uU0pdLSLDgb8Ak4AW4Eyl1IYy+h6MYnH8eY/JaWkgYZJ6TIutpzvpJEmjr9QjFUo9FXjBKu0N7PpEEBgLdcqGWhBGWmYdZx+KeOGbndM3JsMk9QAccQnssA/8ei4c9T1Y8FtY/xZc9E5hG35saIGr94JDvgGHXQiXlpCO8jt/8tVw71e738/7G/z2SFQmfPerL0HzcMrlKzcu4OP7T+CYnccUfFao8dvG8deZxk/lXrkK7TytjFJpCzvK/AwgBXxDKbU7cCDwHyKyO3AR8JBSagrwkPc+OoJKPaA1e4Phx5SFMt1JZ8lwztpKPUWr5Bc4Do5yaxI3HYZEo1RXDrqK8E3ZAIifG+u3YSfZfVMIIIX4tgEln+DAZ/AxE+5Z0F6F/crepmGHmZwW2zh+6H3hnFBM6qm87aCUOup7ichmQ7kATYbyLpRSq9ApnVFKbRGR14EdgQ8Dh3vVbgIeAb4dvMuWiMD2DfDuM7nl7y8sSLKWliTDNyyEd5tzyie1LmPkhgHw7uruwo3v0KmijOqpROrR2SKtoiLKSNIWFmHF8YdhJkxenVLKnKun4Mt5nzkJK8NtpEtGCmD4MZwzyX7m9pykdn4qNPx++8Pk2dqe0/UW+hlKdk7ftqs7gatUWuZQ8vGIyCRgH+AZYIx3UwBYjZaCTN85DzgPYMKECtICiAOL74OWxwsjLHY5IudtS//dmfbGtdByXU75ma2dJFcILMndXe8mT2G037lQy6UX3XQJj98wZuGlbKgFYeQpsV1UvWhf8soyN1DfOH4/nKwJghUb/tLbNSa7S/gZ/kTu+wooZszy31tlDzU1UkMcB/7z7oU0N+QeyyN3G82Zs3eqUa/Ko/KjXgIRGQjcCXxNKbU5+66plFIiYjwTlFLXAdcBzJ49u3yzkDGAR34H9v9c0aq/3fFyjt5tNB/eO3eg6ud/W8jkUQOZd9CknPL/u+ZJ9o/K46/kfC+Vl8e4wQRCukaDuyEsvajs1tb1w7S+cteC7X7J1LP6kNtYGFJP8BuGMVrSz+OXMKWewnLjBK4y0jLXj9mH//34Pqze1JZT9vw7G3hkyZrAht/vKabepJ6K8AaA7wT+pJT6q1f8voiMU0qtEpFxwJoo+9D1iDyo9GIcfgfFT4pQqsii4pWmbCjSn5Io1zMYPieZKToh4/HXSOOvOFcP4QyambxStytU1HL/iGMl1RixMMxGvb3A8Ot+qESD3lsVDowU2xv5n9WR814W+0wYZixf/lKrVTsmyajaY2shDIeZEf1LrgdeV0r9LOuje+hevH0e8Leo+qA74v3EAIYf/A+AyTClXeW/kHedDO4GPqEyGn+E3fLdNFBpoJzrhuPxm7xSnQcogMdfsK/FSqoxYuWRG7aRv2ZEpWMORsxOUaXhnNTZBC4TjmPO5upHsaq9xeOfi54H8GrWoi2XAFcAt3mLurwDnBlhH+ja1YOMQwkFNY1h2j4pBYrqyhXm408mHO5fuJpFq3LH1vslE/zm0/syoF+RQ9c1gcuna2CQJfQErlrgiLBuawdrNuc+Rg9qaqB/YzADpUP/QrH8BU933Q9CJTz+grkRUrmBtZRiCp/k8rZf6RNIfvM+u8P3zO/BUo8JncfL7jo3RlMDz7z9Aem8tvo1OAXScxhEZviVUk/gf9yOimq7BbSu1/8Hljb8fidlsSUGo4rqOW3WjkwdM6ig/It/ep4NrR2BDD/YaPyCqNpE9ewyagD/dffCnLLOtMvu4wZzy+fMayTkU6lUlMEk3QXW+AuQygdPLQy0UW/Pv/F09SecA11ct8+L6qly5Eo1cAyOQlF8qn5oj7E89Pr7LGhZn1M+sCnZswx/3bDNG0IwrI+bj9/sT78T1i2p8Zdv+Jsbk+y/c+HEmqaGROkHCS+qxz/awhBPLQ5COPlubPnx6XsVlD3/zgYu+/uiwG0UHW+xwDTIFljjN4WxZBvecp4AMxp8gO8aJbP8G09XtFd4BtgcnWHYHbaDu1Rf+7bFVuoB8yl03IyxHDdjbDidCkBkGn/dsHVt4Kp+s+r04G5h/bRSJPz2YAiLrZtIOFLwOFhAqfV1TUiiZguxmCgnxDO8cM78wd1yPX5CCZcMitnjT5rfhyQo20g9tikPeozU0wNXFev9hn/YJKvqfilTzVE9RTySiAZ3EyKkS51oXlRPMY3fGMdfowlcJmyjmrpy6YSw3fy8QcolMzBieXcJQeqx2prhSc5X6gmPomNJee9tI9Xq5Xz0I1FGKHJPSNnQ8zn2v+GSlYGq+nkYpthuqE1Uj98U+RxUqQlcpoYdHFTNJnHl41hO79cpG8II5yzcriKAxz9wLOx6DIzLkq1E7I9DuYzerdvj32n/rA/y+huy4feTQU1rINvKNvU2c9eEI5R+As+iXn5S7zf8iSQ0DihdD//HsGIZG3115Yhm7mqpp0SlrLTMJu/CeLF6nmGtErXlYyv1lCHCFNluGXH831wCk+bCoRfCJau6y7vqS2Xua7HvfrsFTspKlHvy/8LXF5m/5xl+FZIr7ee9midwWUo9RdqvF8LS+KtN7x/ctcBv/pJgPriuq/znv0Tk8TsSQON3SydpK6ArCqhODD922UndkAThpCP89vG3uOfl7qfEMW1vMVz6BdP4HQcaM7mesutG6Or11xOLum7oTsK4+EpX/0LuTlCpp1hd33brwEgWwzacs148/tjwZ6GPiWlw1yecUxVZ57XCtMx+BBpMynj8YLxwjNEVnuF36uTEFLHL0a+KHQsLvnb0VJau2ZpTdtgtH2LruIOA4bpjB3weZgdYrazKrp35uEbcB7/BXVNUT50b8XKwDuekPu5lseHPwtfj9zG2OqqnyoO7TnDDb0WXx1+FtWIDYBsBAuFcUDsM7c8OQwu95YGyvfvuf/yPqtgjC4z7rIhjEs4mzSuW+aRlBrP+b0LPmK5v7Gfu1odnFRv+PIxBPWKeIOS7cDIQ1dKLgQaTvMFdP/XDqJ93Gf76ODG192oh9bgRT+93GrCK6mlohjG769fNI2D8fjD3q/DWo/bbHjoRJh+pXx98ASy4HqYeC6Omwpb3u6rlRM00DtQlA0bB6D1gzB7d52O/wWw7+GJW3v9jptv3pgD/xGM++n/Q3WgbRFUDrMM56+Q3xYY/B7MnonX1QiNei6ieQB5GV5I2m4a9wd06ieW3vTgU4Ug9viQa7ETn/8wa4L3wLf1/xGTY9xz7bX/tle7XR39P/xnIOXeTjXCpt0zjF/+dW/Hi5bRtaeNT80fznH1v/LeZhf/Tc6SjHVWnnHDOeqD3R/VYUGyQynRwi2v80Q3uBtf4LbzgehvctdROw4rj98VJUu8uqHF+hh9+A0BlbNPYvMWArx89IapHhNJRdln4jSNWm9jjz8LPn2tMOtz01Ds8+kbuLOCt7Skakj73zggncJWMIii5EIspmZc3uFsn/pjf3Ak/ik6mC6VDyZ4RZhIQnQ4hpLYsjpN0JTwMoPHX930W0GNuPWG+QT6x4c/CT7P/9JyJHDR5ZEF5/0aHgX7J0iJK2eA4BJi5q7wxBrPHZJ6dXGdSD/Yx32GkZfYl0UhmEft6xSYJWljpEGxSF4H9ca13AoVXZ1EvN7PY8Gfh5wU1NyaZOX5I4QfFiGgCl44bLlFJuSXPrgIvpQ6lHhtPSt+0I1Au172p/yeS0Jminj1+K6mHcAyQNuTmmbt+BO2jlVRZI8oJ56wHYo0/j9A0xXKyjAUgWDhnd3ZOvwlphYW6rFY5+Quxl3pC9/jbNsEv99Wv+w+j7jV+C3msa3GZEPDV8w37yv6ppH73N+hgC7vLvD5CVGPDn0WodjqyXD3Bk7RZNoyLUzcev2MZ/RGJd5hq1//3Px/6Da57jd9GRgnrXPeN6sEnlNjihh7mOERUOEGuxzoklnqyCCnQQRPZUQVT4gAAH5BJREFU4C7Gwd0tbZ1sbksBMHh7B9Lhsm5ru29InQkHt340frGM449icDezLxIN3nhNfXv8WMhjUQ/u+najjndfOdiGc0YehBCQ2PBnocIMCRQH3FRYrXWR8InjP/uGZ1m+fjuNCeELqXdpUm38fMlijt69cOWxhoTw1rptHHfVYznlDwA7rnkM2L/gO9XGenA3CpucGZxPNEJqS917/GDj8YeXxtr3M5/B3aD0hGgZ29Qi9UJs+PMI7W7sNGitPWTEJ4pgS1uKWz53gF6u8R8Pg1KccZx5hcvJowZy/1cOpjOd286mf32U3YfWT8oG2wUuQteDlaujnZyEdyzr2+O3+f2heZ4+kpzf8bAZ+tJPJfW7vyEzodLmPK0P1yE2/HmEdlCcZDQev48E0taZpinp6fpuuuhSkyLCrqML1/Nl5I7Qtr6wvAaUF8cfcicy6xpkjmWde/xSo1hJ0/noP4HLLutqvZOwDOesF2LDn0WoT5ZOoqpx/O0pl34N3li9m/JPy1uM5hHwyl/gzs/BkBILPG9aoetsWwvpFAweV6Jx0TcjcSDVlttG/mtgyPZOzu9YCQvehv28TJjvPQeL7+tucstqGDASnCR7runPyl1Ot/u9GTrb4MmruvuVwU3r49h1LOvd47eMmAnF4fdvxDdlQ14XWztSnHvjc7Slcq+XTds72XlEsLU0aoXjk8fLj6L5vapIbPizCDUyJCKP32/CSHuOx58qb6WlaSfACzfBq7fBrHmQ7Geut+oVWP407HEavPZXXbbbyTCoiPFffD9sfk+/Hr07DJ0Abzygk481NGuDPukQGL2b/p2pFGPUOnjo+92G/+Vb4b0F3StMvfxnfSPZ/cMctuJxbt3lDPvfDLDhbXj8p7l5dFrXw8I7dN8k0WM8/qrH8ftKPUXSOeS939DayRvvb+G6s2cX1B0/rAwHpoo48czdnk+occNRST0+ccNt+R5/OYZ/1FSYMAfWvwXHXg79DHIQwFPXaMN/yDe6Df9BX8lb8i+P9W91G/5Jh8CUY7Th3+fTMHCMNvx7fRz2+RQAmzZu54ZFt/Eh98XuNlRaf+/I/9Lvn70Oxs6EQ76Bs+Rf5RsyN62zWZ7wk+6yNYu14ZeEdyw9b7Qe3DUfjGvu+hDqmIhFVI+eZJb7YSrt0twvwb4Th4XTnypiG85ZLyGqseHPIlSd2ElCujOkxrKaNXj8Sik6Ui6NiWzDbxnHn0+xWbCZz7LrlJo1m183s6PFMbbnCKRxcgfI/XIQOQ0kVKr8gUDTGsWZ907P0vgDD5yGFcdv+aFpTePOtEtDomdOKUqI0N7p8o/XVhd8tu/EYYwc6PPUXGNiw59FqAs/RKXxGzyM9pRLY9LpXmy8XI8/Gyly48h8ln1zKWn4s+o6idw2Mu1k1RGENJK7D1Xa3K9EA45KlX/sMlp+Nl198gz/yhf0hK56cNeKEFjjJ0ypxyz2GKN6DDU704oG3zVM65sB/RIcu8cY7nj+vZzyt9dt44SZ47jgmKk55fUyGzk2/PnUucafcOBPT7/DY1mZQlNpRVN2llA3VTSqJxDFnhgyF2m+MQ/anji5xt5wIxGBtMqbBKcUxkWOnSSOSpWfsiGT1C4byfqNU4+F7esBgUkHl7mR6LEx5GHOWbGdwJX/WWfapSFZe2NYDsmEw1Uf36eg/Oan32Hhik016FEwYsOfRbhRPdEY/i8cviuvGk6osw+a2P2mah6/U1jm+52sugUef6HUI0BKmaQes8efrMTjV0U8ficBwybC4ReV23oVsU2HULmxtY7qyWw8i860ItlDPX4/xgxu4qZ/t3DNv97MKX93fWtdqIWx4c9CK7j1Pbi788gB7DyyRIibSbqwpdiFWI7Gn+PxJ3LbMLQn4kk9yu0efDEZaPCWRQSn3PWCXUO6ZdNvrHNsVrcKMy2zTVSP1vhzv9GZzhqf6iXMnjiMY/cYy9b2XBtwwsyxzNjBMtNvBMSGP5tQB3ej0fgDEYbHXwyDJl/yRpN/k8jW0H2kHld5ByMzg9ZvcDehf2uiXMNvGjuQrP71ELQ3bRPPGVVP7NIyp9KKZKIO3OAQGTagkW8eO63W3fAlNvxZhDu4G43HH4ioDb9xcLeU4bcd3IVUJnmscoFE9w0gH8/jT6gy97cpm2mlT0w1QAQeX7qO7Z25N8BRg/px6j7j82qHlKunSCSRf1rmXDrdnhvV01OJ93YWoWbOSzSAG344ZyDSnREb/kwoZrYxtwnnzNb1/QZ3hVTG4888OflJPd5AtlOu4TeNHXS97zme6El77sDEEc2s3dLe9bdyYxuX/HVhQd0wz3XjQiz4SD0UPg10plwaepnHX+/EHn9U1NTjT0dr+DNYxfFnXdhOnq7vF8evMh6/Z/hdl+1p+PwNz9KZdrkFHTb3Xzc8z5+Axoqknvw4/kx/e86szAN3GcGBu4zIKWvrTHPLs+8W1A3vV/nk47d4Cki5Kvb4q0y8t7MIdVZdb9b4M9hIPdnke/nGqB7plnqyPP6t7S5LVm/hP47YFdAyxhePnIoSh2Om5xq9wLi9Q+ox4ZfiI9TBXV8jbyjDIPWkXZKx4a8qkVkHEbkBOAlYo5Sa4ZUNB/4CTAJagDOVUhui6oMtoebj780af4b8EE2b75WI40fAzRyNLo8/jRJhQL8Ec3cdCcDAfkn92mmgOVFmYnTlGjz+3mH4E46P4Q8xL5UxqkeZJ3A5jvDlW16kMWveyfub29ijDiJd+hJRWocbgV8Cf8gquwh4SCl1hYhc5L3/doR9sCJUj7+hGT54E34yJaQGLWj9oLzsnLaU6/HnDO7m3QQyzYkXx69FYV2oXNIkSJhmaiUayr/RFovj7+F0TeZ2VffMbsKbQRo0EVuGG+btx7pt7QXlU8f45IWKiYTIDL9S6jERmZRX/GHgcO/1TcAj1JPhD3PtxfH7weefrM0AryS6slxGux0LjT/ne6Xj+HVqCtGHI0vqcXFwTHfnSnIjtW8t1Ct6UBhnMUREe/1K4ZBn+ENK2VA0I1seM8fHnn09UO3B3TFKqVXe69VA4bqAvQURGLN7rXsRLTZx/Nnk6PpZ3n+WJcoN5+yWenwNfyVRVHedV1iW6VMPTLmbT2axkIZqPsTU99IFfZ6auTVKx3T5XlUicp6IPCciz61du9avWrh9Ij5ZrbBJ0paNiDmOP6+KKgjndHERs9TjNOgFYcKil0g9oO+x+csDao0/nJQNfjN3Y+qXahv+90VkHID3f41fRaXUdUqp2Uqp2aNGjapO78Ic3O0LlOvxSyJvcLfwNBS8tUwlUejxGzX+pPb4U4X6cVl0zVXo+WeEaXnAqkT1hNB+TDRU2/DfA8zzXs8D/lbl7RelJyzuXFdk76uSHn92HL+Pxp/XtMrUzWToVC4uDsa5Pk6D1vgvGw0v3RKs/4/8CC4toTn3AqnHcQTXEPAU5amulxiMr6V6JTLDLyJ/Bp4CponIeyJyLnAFcIyILAWO9t7XDWF5QX2GHMNfZhy/iK/Uo+/EiZzB3TSOf1RP53b9evPKYP1oeTx4n3swSUdI5Vn+MBdiMS62Hk7zMRERZVTPJ3w+OiqqbVZKqGvu9jVKSj1ZpiBf1zfcNLR2rHI9flcbfnNUTwN0tgbsS2YjfeNgZ6J6sgnrXDetqNX1WYnvdnR0sGzZMlpbWyvvSB+nubmZyZMn09jYGKh+nLIhJhysPP48ecfH43eVVzc7nFP5efzJbsMftC9BBqQbmoK1Vcc4Uij1RL0SVJBw0WXLljF06FCmTZuG08vy8VcT13VZvXo1CxcuZObMmTQ0lF6EKd7bWdTLsmg9kpLuY954QIk5AI6IlhAkaxUupXDFx/CX5fEHMfwl1j7oAZg8fgjvgce4AlcAsae1tZUxY8bERr9CHMdh7NixuK7L/Pnzi6bE7vpOFfrVYwh1sfW+hs2OKzD8Jqkne3C3O6onrcQ/jr+jNevbAftRisbmYG3VMdrjz5d6whvP8pd6Sm8hNvrh4DgOIsKbb75Je3vpyLZ4r2cRD0hFicp9nZ+tM4+uMMGgg7tOsntwVwXN2RPA9DX2Eo+/IJwzHC/Hr4nYiaoNIkIqVXo+S2z4s9AhaLXuRY0ZNK50nf7D7NttzsqcmWjs9vKT/brLs4xsJhRQJRrh9nPgVwfBhhZS0pC7qPrQCV6bDdC5Tb9WAbOiBvH4M+33YJLGwd2Q4viRQNJCT+eEE05g48aNRet897vf5cEHHyyr/UceeYSTTjqprO+WQzy4G5PL4RfBvvOK15nyIfjyC/r1BYt1UrhSHPZt2PNjkO6AsTO1wf/yC92G9csvwIjJOV8RAfWJW5Gt7+sCJ8HGbbvivPt+97abvDh8JyucM2g67OyxgHPuL/z880/CkPyVq3oeOo7fMIEryjj+6JquKkoplFLcf7/h/MjjBz/4QRV6FA6xx59Hnx/cTTSU9nJFuo304HEwdkbpdpsGw7g9Yfzsbi8/29DnGX3wdP4xM2DK0fpv8hGkSHZLPYPHdWvw2VE9QaWejMefaIQhOxZ+PnYG9B8arK06JiGmwd1wTLPvzaMHST0/+9nPmDFjBjNmzOCqq66ipaWFadOmcfbZZzNjxgyWL1/OpEmTWLduHQA//OEPmTZtGgcffDCf+MQnuPLKKwE455xzuOOOOwCYNGkS3/ve95g1axYzZ85k8eLFADz77LPMmTOHffbZh4MOOoglS5bU5Df3ao//zuff48Z/txSU929I8Idz96cpL2tVrEvWF5KJ7Mm6GaeVMqdsSDZB22b9OrDh99pJd+jv91IcR0ilI0rZQHiTwSZd9PdwGsqi5YoTi37+/PPP8/vf/55nnnkGpRQHHHAAhx12GEuXLuWmm27iwAMPzKm/YMEC7rzzTl5++WU6OzuZNWsW++67r7HtkSNH8sILL/CrX/2KK6+8kt/97ndMnz6dxx9/nGQyyYMPPsgll1zCnXfeGdrvDUqvNvyHTB3JlDEDC8rPvek51m5pZ6fhuREb8QSu+kLwYvmzcF1FwnSQBoyG5c94lWw1fskdg+hlJAxJ2iDa9CQK80IsxShlpKPgiSee4NRTT2XAAD2+dNppp/H4448zceLEAqMP8OSTT/LhD3+YpqYmmpqaOPnkk33bPu200wDYd999+etf/wrApk2bmDdvHkuXLkVE6Oyszbrcvdrwjx7UxOhBhZ7c8OZGtrYXjnzHcfz1hSNSEA+edpU5qmfgaFjxnH5tK/U0j+hasL03YkzSRohJ2kyLrffwp+fMjaAS+vXTkmYikeiKtPnOd77DEUccwV133UVLSwuHH354xdsphz6p8Q9qSpoNfw36ElMEQ+bHtFLmOP4Ru3a/DhrV02+w/j/5yPL610Po15Dga395ieOueqzr7xu3vRzKAudajguhkzXikEMO4e6776a1tZVt27Zx1113ccghh/jWnzt3Lvfeey9tbW1s3bqV++67z2p7mzZtYscd9XjSjTfeWEnXK6JXe/x+DGxKsqXN/IjVk72U3obpULiuwmivJh6UVSmg4U82wdyvwTHfL6d7PYZrzprF2i2Fk3pGDepnqB0OYU4Qi5JZs2ZxzjnnsP/++wPw2c9+lmHD/MOV99tvP0455RT23HNPxowZw8yZMxkyJPiqYhdeeCHz5s3jsssu48QTqy9tZeiThn/s4CY+f/MLJPMkg/aUS/+qLlMUUwydrydP6lE+Uk+2VBNU6qnWovQ1ZtSgfpEZ+a4Z1nn0pLTMF1xwARdccEFO2cKFC3Pet7S0dL3+5je/yaWXXkprayuHHnpo1+ButgefXX/27Nk88sgjAMyZM4c33nij67PLLrsMgMMPP7yqsk/vP+sNXH7qTL5zUuGyiAlHCiJ9YmqHI8JP57+RI0m8tnITE4Yb0igksrISBjX8Kt0nDH+k9AzbHirnnXceixYtoq2tjXnz5jFr1qxad8maPnnWJxxhQL8++dN7FN8/ZQ/Wbs2VKOZMHsHBu44srOxkefxBpR431auWWKwV5iRtvfeecMstARf6qWNi6xdTt5wxe6fglbPz/QQd3HXd2PBXiCC8tHwDtzzzbk55T4/q6e3Ehj+m92Gj8dusIxBTwOHTRrFpewevrsjNY3P+YbvUqEcxQYgNf0zvI7DU0xlr/BVy4C4jOHCX3jv5rbfSJ+P4Y3o5QaWeVEdudtCYmD5CbPhjeh9BpZ50e2z4Y0IjO7XyPffcwxVXXOFbd+PGjfzqV7+y3sall17alRSuEmLDH9P7yF9g1o9Ue69OzhYTDul0wCfILE455RQuuugi38/LNfxhERv+mN5HYKmnPTf+P6Z2KAXtW8L/K5FPoqWlhenTp/PJT36S3XbbjdNPP53W1lYmTZrEt7/9bWbNmsXtt9/O/PnzmTNnDrNmzeKMM85g69atADzwwANMnz6dWbNmdSViAz2Z60tf+hIA77//Pqeeeip77bUXe+21F//+97+56KKLWLZsGXvvvTff+ta3APjJT37Cfvvtx5577sn3vve9rrYuv/xypk6dysEHHxxaGud4ZCum93DWbfDeAtjQEqx+Ovb464aOrfA/ESx6c/F70G9Q0SpLlizh+uuvZ+7cuXzmM5/p8sRHjBjBCy+8wLp16zjttNN48MEHGTBgAD/60Y/42c9+xoUXXsjnPvc5Hn74YXbddVc+9rGPGdv/yle+wmGHHcZdd91FOp1m69atXHHFFSxcuJCXXnoJgPnz57N06VKeffZZlFKccsopPPbYYwwYMIBbb72Vl156iVQqVTQNtA2x4Y/pPUw9FtYt1WGaQUi1QzL2+OuCxoHaSEfRbgl22mkn5s6dC8CnPvUp/vd//xegy5A//fTTLFq0qKtOR0cHc+bMYfHixey8885MmTKl67vXXXddQfsPP/wwf/jDHwCdqXPIkCFs2LAhp878+fOZP38+++yzDwBbt25l6dKlbNmyhVNPPZXmZj1b/ZRTTrHeBSZiwx/Tu2gaAtuLr43aRaodEvHgbl0gUtIzj27TYnyfSc2slOKYY47hz3/+c069jLceBkopLr74Ys4///yc8quuuiq0bWQTa/wxvYsBo2DbumB1e/nKWzHBePfdd3nqqacAnY7h4IMPzvn8wAMP5Mknn+TNN98EYNu2bbzxxhtMnz6dlpYWli1bBlBwY8hw1FFHce211wJ6oHjTpk0MGjSILVu2dNU59thjueGGG7rGDlasWMGaNWs49NBDufvuu9m+fTtbtmzh3nvvDeU3xx5/TO9i4ChYvwz+en7putvWxVJPDNOmTeOaa67hM5/5DLvvvjtf+MIX+MUvftH1+ahRo7jxxhv5xCc+QXu7zh112WWXMXXqVK677jpOPPFEmpubOeSQQ3KMeYarr76a8847j+uvv55EIsG1117LnDlzmDt3LjNmzOD444/nJz/5Ca+//jpz5swBYODAgdx8883MmjWLj33sY+y1116MHj2a/fbbL5TfLOr/t3f+QVqVVRz/fAGBENAlHQfFWkTTAX+Bq7Fq5SiYGJkVDTKWP6CafkyaUzqaFdFoM2allI05Y4qjjToqKUOTqPgjhhphl+SHKO4yrgoJmmNUKLmwpz+e88LdZWV5l3Xfvfc9n5k7773ned7nnnPP+5773Ofee54czKJQV1dnDQ0NlVYjyAPb34Plt6cbt13RfxCc/NUI/hWisbGxR25U7gstLS1MnTp1tzTMeaSxsZGlS5cyc+ZMhg5N9zYkNZpZXce60eMPisWAgVD/rUprEQR9mhjjD4KgaqmtrS1Eb79cIvAHQVAx2vb2Letgj5R7HCPwB0FQEYYMGcLmzZsj+O8jbW1tbNq0idbWzucR74wY4w+CoCKMGTOGpqYmNm7cmJv5efsqra2tNDc3069fPwYP7voR5Qj8QRBUhIEDBzJu3DhWrFjBkiVLIvj3AFOmTGHAgK7DegT+IAgqyvjx4xk9ejRbt26ttCq5RRLDhg1j+PDhe1U/An8QBBVFEjU1NdTU1FRalaohbu4GQRBUGbl4c1fSm8Ar3fz6QcBeJm/JHUW2DYptX9iWX/Jk30fN7OCOwlwE/n1BUkNnrywXgSLbBsW2L2zLL0WwL4Z6giAIqowI/EEQBFVGNQT+3afEKQ5Ftg2KbV/Yll9yb1/hx/iDIAiC9lRDjz8IgiDIEIE/CIKgyih04Jd0jqR1kpolXV1pfcpF0uGSnpK0VtLzki53+QhJj0tq8s8al0vSr93eVZImVNaCrpHUX9LfJS307dGSnnUb7pc00OWDfLvZy2srqXdXSDpQ0oOSXpT0gqT6gvntCv9NrpF0r6TBefWdpDskvSFpTUZWtq8kXez1myRdXAlb9pbCBn5J/YHfAlOAscAMSWMrq1XZbAe+Z2ZjgYnAt92Gq4HFZnYUsNi3Idl6lC9fB27tfZXL5nLghcz2DcBNZnYk8DYwy+WzgLddfpPX68vMBR41s2OAE0g2FsJvkg4DLgPqzOxYoD9wAfn13TzgnA6ysnwlaQQwG/g4cAowu3Sy6JOYWSEXoB5YlNm+Brim0nrto02PAJOBdcBIl40E1vn6bcCMTP2d9friAowi/anOBBYCIr0ROaCjD4FFQL2vD/B6qrQN72PXAcDLHfUrkN8OA14DRrgvFgKfzrPvgFpgTXd9BcwAbsvI29Xra0the/zs+nGW2OCyXOKXx+OBZ4FDzOx1L9oEHOLrebP5ZuAqoDQTx4eBf5nZdt/O6r/TNi/f4vX7IqOBN4E7fRjrdkn7UxC/mdlG4BfAq8DrJF80UgzflSjXV7nyYZEDf2GQNBR4CPiumf07W2ape5G7Z3IlTQXeMLPGSuvyATAAmADcambjga3sGioA8us3AB/C+BzpBHcosD+7D5UUhjz76v0ocuDfCBye2R7lslwhaT9S0P+Dmc138WZJI718JPCGy/Nk82nAeZJagPtIwz1zgQMlldKFZ/XfaZuXHwC81ZsKl8EGYIOZPevbD5JOBEXwG8Ak4GUze9PMWoH5JH8WwXclyvVVrnxY5MC/HDjKnzQYSLr5tKDCOpWFJAG/B14ws19lihYApacGLiaN/ZfkF/mTBxOBLZnL1T6FmV1jZqPMrJbkmyfN7ELgKWCaV+toW8nmaV6/T/bCzGwT8Jqko110FrCWAvjNeRWYKGmI/0ZL9uXedxnK9dUi4GxJNX5FdLbL+iaVvsnwQS7AucBLwHrg2krr0w39TyddYq4CnvPlXNL46GKgCXgCGOH1RXqSaT2wmvTURcXt2As7zwAW+voRwDKgGXgAGOTywb7d7OVHVFrvLmw6EWhw3z0M1BTJb8Ac4EVgDXA3MCivvgPuJd2raCVdrc3qjq+AmW5jM3Bppe3a0xIpG4IgCKqMIg/1BEEQBJ0QgT8IgqDKiMAfBEFQZUTgD4IgqDIi8AdBEFQZEfiDXkfSKEmPeBbD9ZLmZjI5XiLplj6g4/nZpH6SfippUg+0e4Ykk/TZjGyhpDP2tW1vq0XSQT3RVlBcIvAHvYq/8DMfeNhS5sOPAUOB6z/AfQ7outZunE/K6gqAmf3YzJ7oIZU2ANf2UFs9RjePU5BDIvAHvc2ZwDYzuxPAzHYAVwAzJQ3xOodLetqvCGYDSNpf0p8krfQc8NNdfpKkZyQ1SlqUec3+aUk3S2oArpX0iqR+mbZek7SfpK9JWu7tPuRvo54KnAfcKOk5SWMkzZM0zb9/lidfW62Uy32Qy1skzZG0wsuOeZ9jsBLYImlyx4Jsj11SnaSnff0nku6StMRt+YKkn/t+HvXUHiWucvkySUf69w92+5b7clqm3bslLSW9iBVUARH4g95mHCmT404sJZ57FTjSRacAXwSOB74kqY6UBOwfZnaCpRzwpWD3G2CamZ0E3EH7K4eBZlZnZnNIbz1/yuVTSSmDW4H5ZnaymZVy5s8ys7+SXs2/0sxONLP1pQYlDSblb59uZseRErJ9M7PPf5rZBFKe9u/v4ThcD/ywq4PVgTGkE+d5wD3AU67Du8BnMvW2uPwWUgZUSHmQbjKzk0nH9vZM/bHAJDObUaY+QU6JwB/0RR43s7fM7F3SsNDppNfjJ0u6QdInzGwLcDRwLPC4pOdIgXRUpp37O6xP9/ULMmXHei96NXAh6cS0J44mJSh7ybfvAj6ZKS8l0msk5XjvFDP7C4Ck07vYX5Y/+8lqNWnyk0ddvrrDvu7NfNb7+iTgFj9OC4DhSllfARb4sQ6qhBjTC3qbtexK5AWApOHAR0g5TiawewpcM7OXlKa5Oxe4TtJi4I/A82ZWT+dszawvAH6mNFPSScCTLp8HnG9mKyVdQsobtC/8zz930PX/q9Tr356RbWdXh2xwZ22bWZukVtuVb6Wtw76sk/V+wEQz25ZtMN1yaXecgiogevxBb7MYGCLpItg5ReYvgXlm9o7Xmaw05+mHSDdZl0o6FHjHzO4BbiSdINYBB0uq97b2k9Rpj93M/kvK2DqXlBBuhxcNA173YaMLM1/5j5d1ZB1QWxo7B74CPFP2UUg6PUZK3nZ8RtxCOjFBGpLpDtMzn3/z9ceA75QqSDqxm20HBSACf9CreC/186Sx+yZS9tRtwA8y1ZaR5iBYBTxkZg3AccAyH6qYDVxnZu+Rrh5ukLSSNI5/6h52fz/wZdoPAf2INKvZUlK2yRL3AVf6TdwxGf23AZcCD/jwUBvwu/KOQjuup30e9znAXL8pvaPzr3RJjaRVpPmMr3DZZUCd0gTha4FvdFfhIP9Eds4gCIIqI3r8QRAEVUYE/iAIgiojAn8QBEGVEYE/CIKgyojAHwRBUGVE4A+CIKgyIvAHQRBUGf8HrifmM7vuPPMAAAAASUVORK5CYII=\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "YUKe5QbBXyyI"
      },
      "source": [
        "pred.clip(upper=pd.Series({'dist_err': 38}), axis=1,inplace=True)"
      ],
      "execution_count": 53,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "7NH6uwDDlCPT"
      },
      "source": [
        "Results"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "gNorICFWgdb7",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "78919385-e1ea-43a6-db31-58b5eed681b6"
      },
      "source": [
        "knnmean=pred['dist_err'].mean()\n",
        "knnmax=pred['dist_err'].max()\n",
        "print(\"MEAN DISTANCE ERROR USING KNN REGRESSOR: \", knnmean,'m')\n",
        "print(\"MAXIMUM DISTANCE ERROR USING KNN REGRESSOR: \",knnmax,'m')"
      ],
      "execution_count": 54,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "MEAN DISTANCE ERROR USING KNN REGRESSOR:  19.224908935684294 m\n",
            "MAXIMUM DISTANCE ERROR USING KNN REGRESSOR:  38.0 m\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "SSY1snXmvv9H",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "626dcd94-b59f-449f-9c57-0625b408ede6"
      },
      "source": [
        "count = len([i for i in pred['dist_err'] if i  < 2])\n",
        "low_error = count / len(pred['dist_err']) * 100\n",
        "print(\"Percentage of data with less than 2m error : \",low_error,'%')"
      ],
      "execution_count": 55,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Percentage of data with less than 2m error :  0.0 %\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "azh4uTEOr-ZG"
      },
      "source": [
        "##Summary\n",
        "\n",
        "We discovered that the data set downloaded was in relatively clean state and therefore we did not have to clean up the data. However, feature engineering was required to come to the conclusions regarding the data. \n",
        "\n",
        "Since the number of Reference point in the training dataset is limited to 13, and is not uniformly distributed over the localisation area, satisfactory results could not be obtained using RSSI fingerprinting technique.\n",
        "\n",
        "Moreover as RSSI values at the same time instant were sometimes available to 3, sometimes to 4 - estimating the location from multilateration was not possible.\n",
        "\n",
        "The best result was achieved using AdaBoost Regressor.\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "B_V-Xj2xgU7Q"
      },
      "source": [
        "## Challenges of using RSSI in Practical Scenario:\n",
        "\n",
        "RSSI values are highly sensitive to changes in the environment and they vary rapidly, reducing its reliability.\n",
        "\n",
        "One study by Parameswaran et al. illustrates the same unreliability -\n",
        "\n",
        "When RSSI was measured from the North of an AP the RSSI vs distance plot obtained is illustrated below:\n",
        "\n",
        "![picture](https://drive.google.com/uc?export=view&id=1qNSJSlHc3YwT3Z36jCHXKiw8SVN6oE4T)\n",
        "\n",
        "The same AP when measuring RSSI from the east yielded this RSSI vs distance plot:\n",
        "\n",
        "\n",
        "![picture](https://drive.google.com/uc?export=view&id=1Sapha1-WAIaVeyihs9S4a7sVap3ptlqj)\n",
        "\n",
        "This clearly illustrates the unreliability of RSSI when using it for localization."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "KqFkQL2dsw9N"
      },
      "source": [
        "## Future aspects for improvement\n",
        "\n",
        "- Use of Link Quality Index, Time of Arrival and Angle of Arrival can be done instead of RSSI for more robust localiztion results, as these parameters are not so heavily influenced by the environment as RSSI.\n",
        "\n",
        "\n",
        "- Filters such as Kalman Filter and Gaussian Filter can be applied to the data if the data collected is in the form of a Time-Series with changes in Latitude and Longitude. \n",
        "\n",
        "- More number of reference points in the data provided would significantly improve the performance of the models."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "TGW7u0naspvV"
      },
      "source": [
        "## References\n",
        "\n",
        "- Shen, Y., Hwang, B. and Jeong, J.P., 2020. Particle Filtering-Based Indoor Positioning System for Beacon Tag Tracking. IEEE Access, 8, pp.226445-226460.\n",
        "- Li, G., Geng, E., Ye, Z., Xu, Y., Lin, J. and Pang, Y., 2018. Indoor positioning algorithm based on the improved RSSI distance model. Sensors, 18(9), p.2820.\n",
        "- Bullmann, M., Fetzer, T., Ebner, F., Ebner, M., Deinzer, F. and Grzegorzek, M., 2020. Comparison of 2.4 GHz WiFi FTM-and RSSI-based indoor positioning methods in realistic scenarios. Sensors, 20(16), p.4515.\n",
        "- Sung, Y., 2016. RSSI-based distance estimation framework using a Kalman filter for sustainable indoor computing environments. Sustainability, 8(11), p.1136.\n",
        "- Parameswaran, A.T., Husain, M.I. and Upadhyaya, S., 2009, September. Is rssi a reliable parameter in sensor localization algorithms: An experimental study. In Field failure data analysis workshop (F2DA09) (Vol. 5). IEEE.\n",
        "Li, G., Geng, E., Ye, Z., Xu, Y., Lin, J., Pang, Y.: Indoor Positioning Algorithm Based on the Improved RSSI Distance Model. Sensors. 18, 2820 (2018).\n",
        "- Kraus and Fleisch, Electromagnetics, 5th Ed., McGraw-Hill, 1999\n",
        "Navarro, E., Peuker, B., Quan, M., Clark, A.C. and Jipson, J., 2010. Wi-fi localization using RSSI fingerprinting (Doctoral dissertation, California Polytechnic State University).\n",
        "- Chan, S. and Sohn, G., 2012. Indoor localization using wi-fi based fingerprinting and trilateration techiques for lbs applications. International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, 38(4), p.C26.\n",
        "- Choraś, Ryszard S. (2010). [Advances in Intelligent and Soft Computing] Image Processing and Communications Challenges 2 Volume 84 || Evaluation of Smoothing Algorithms for a RSSI-Based Device-Free Passive Localisation. , 10.1007/978-3-642-16295-4(Chapter 52), 469–476. doi:10.1007/978-3-642-16295-4_52\n",
        "\n"
      ]
    }
  ]
}