Code-Example-017-SVM.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Support Vector Machine\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "\n",
    "from scipy.stats import norm\n",
    "from sklearn.datasets import make_blobs\n",
    "\n",
    "from matplotlib import pyplot\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "import matplotlib.cm as cm\n",
    "from random import sample"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Generating Some Example Data with 2 Features"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "n_feature = 2 # the number of features\n",
    "n_components = 2 # the number of clusters\n",
    "\n",
    "n = 1000 # the number of total samples"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[  1.9738175 ,  -7.50934555],\n",
       "       [ -1.19505856,  -5.8378053 ],\n",
       "       [  0.37631303, -13.57342824],\n",
       "       ...,\n",
       "       [ -0.8821049 , -19.09837957],\n",
       "       [ -2.82660508, -13.01978618],\n",
       "       [ -0.92520746,  -7.54158259]])"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# generate isotropic Gaussian blobs (data from mixture Gaussian distributions)\n",
    "X, y = make_blobs(n_samples=n, # the number of total samples\n",
    "                  centers=n_components, # the number of clusters\n",
    "                  n_features=n_feature, # the number of features\n",
    "                  cluster_std = [4] * n_components, # the std of each cluster\n",
    "                  random_state=2) # 2\n",
    "\n",
    "# X = np.hstack((np.ones((m, 1)), X))\n",
    "\n",
    "X"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[1 0 0 1 1 0 1 0 1 1 0 0 0 0 1 1 1 0 0 1 1 0 0 0 1 1 1 0 0 1 1 0 0 1 1 0 0\n",
      " 1 1 1 0 0 1 1 0 0 1 0 0 0 0 1 0 0 0 0 1 0 1 0 1 1 1 1 0 0 1 0 0 0 1 1 1 1\n",
      " 1 1 0 0 1 0 0 0 0 0 0 1 1 0 0 0 0 0 1 1 0 0 1 0 0 1 0 1 1 0 1 1 1 0 0 0 1\n",
      " 0 1 1 0 1 1 0 0 0 0 1 1 0 0 0 1 1 1 0 1 1 1 1 0 0 1 0 0 1 0 1 0 1 0 0 1 1\n",
      " 1 1 0 0 1 1 0 0 0 1 0 1 1 0 0 0 1 0 0 1 1 1 1 1 1 0 1 1 1 1 1 0 1 1 1 1 1\n",
      " 0 1 1 1 1 1 0 1 1 1 0 0 0 0 0 1 1 1 1 0 1 0 1 0 1 1 1 1 0 1 1 1 1 1 1 1 1\n",
      " 1 0 1 1 0 1 1 1 0 0 0 0 0 0 1 1 0 1 1 1 1 0 1 0 0 0 0 1 1 0 0 0 1 0 0 1 1\n",
      " 0 0 1 0 0 0 0 0 0 0 0 1 0 1 0 0 1 1 0 0 0 0 1 0 0 0 1 1 1 0 1 0 1 0 1 0 0\n",
      " 0 1 0 0 1 0 0 1 0 1 1 1 1 1 0 0 0 0 0 0 1 0 0 0 1 1 0 0 0 1 0 0 1 1 1 0 1\n",
      " 1 0 0 0 0 1 1 0 0 1 1 0 1 1 1 0 1 0 0 0 0 0 0 1 1 0 0 1 1 0 0 0 1 1 0 0 0\n",
      " 1 0 1 1 0 1 0 1 1 1 1 1 0 1 0 1 1 1 0 0 1 0 1 1 0 1 1 0 1 1 1 1 1 0 0 0 0\n",
      " 1 1 1 0 0 1 0 1 1 1 0 0 0 1 1 0 1 1 0 1 1 0 0 0 1 1 0 1 0 1 1 0 0 0 0 1 0\n",
      " 0 0 0 1 0 1 1 0 1 1 1 1 1 0 0 1 1 0 1 0 0 1 1 0 0 0 0 1 0 0 1 1 0 0 0 0 0\n",
      " 1 0 0 1 1 0 0 0 1 1 0 0 0 1 0 0 0 1 0 1 0 0 1 0 0 1 1 1 0 0 0 0 0 1 1 1 0\n",
      " 1 1 0 0 1 1 1 0 1 1 1 1 1 1 0 1 0 1 1 1 0 0 1 1 1 1 1 0 1 0 1 1 1 1 1 1 1\n",
      " 0 1 0 0 0 0 1 1 1 1 1 1 0 0 1 1 1 0 0 0 0 1 0 1 0 0 1 0 0 1 1 1 0 1 1 1 1\n",
      " 0 1 1 1 0 1 0 0 0 1 0 1 1 1 1 1 0 0 1 0 1 1 0 0 0 1 1 0 0 1 1 0 0 1 0 0 0\n",
      " 0 1 1 0 1 0 1 0 1 0 1 1 1 1 0 1 0 0 1 1 1 0 1 0 1 1 0 1 1 0 0 1 0 1 1 1 0\n",
      " 1 1 1 0 0 1 1 1 1 0 0 1 0 1 1 1 0 0 0 0 0 1 1 0 0 1 0 0 0 0 0 0 0 1 1 0 0\n",
      " 0 0 0 1 1 1 1 0 1 0 0 0 0 0 1 1 0 1 0 1 1 1 0 0 0 0 0 1 1 0 0 0 0 0 0 1 0\n",
      " 0 1 0 1 0 0 1 0 1 0 0 0 0 1 1 0 0 1 0 1 0 0 0 1 1 0 0 0 0 1 0 1 0 0 1 0 0\n",
      " 1 1 1 0 1 1 0 1 1 1 1 1 0 0 0 1 0 0 1 1 1 0 0 1 1 0 0 0 1 1 1 0 1 1 0 1 1\n",
      " 1 1 0 1 1 0 0 1 1 0 0 1 1 1 1 0 1 0 0 1 1 0 0 0 1 1 0 1 1 0 1 0 1 0 1 1 0\n",
      " 0 0 0 1 1 0 0 1 0 0 0 1 0 1 1 1 1 1 1 0 0 0 1 0 0 0 1 1 1 0 0 0 1 1 1 1 1\n",
      " 0 0 1 0 1 1 1 0 0 1 1 0 0 1 1 0 0 1 1 0 1 0 1 1 1 0 0 0 0 0 1 1 1 0 1 0 1\n",
      " 1 0 0 0 1 1 1 1 0 1 0 0 0 0 0 0 1 1 0 0 1 1 1 0 1 0 1 1 1 0 1 1 0 0 0 1 0\n",
      " 1 1 0 1 1 0 0 0 0 1 0 1 1 0 0 0 0 0 0 0 1 0 0 1 0 0 0 1 1 0 1 1 0 0 1 0 0\n",
      " 1]\n",
      "[ 1 -1 -1  1  1 -1  1 -1  1  1 -1 -1 -1 -1  1  1  1 -1 -1  1  1 -1 -1 -1\n",
      "  1  1  1 -1 -1  1  1 -1 -1  1  1 -1 -1  1  1  1 -1 -1  1  1 -1 -1  1 -1\n",
      " -1 -1 -1  1 -1 -1 -1 -1  1 -1  1 -1  1  1  1  1 -1 -1  1 -1 -1 -1  1  1\n",
      "  1  1  1  1 -1 -1  1 -1 -1 -1 -1 -1 -1  1  1 -1 -1 -1 -1 -1  1  1 -1 -1\n",
      "  1 -1 -1  1 -1  1  1 -1  1  1  1 -1 -1 -1  1 -1  1  1 -1  1  1 -1 -1 -1\n",
      " -1  1  1 -1 -1 -1  1  1  1 -1  1  1  1  1 -1 -1  1 -1 -1  1 -1  1 -1  1\n",
      " -1 -1  1  1  1  1 -1 -1  1  1 -1 -1 -1  1 -1  1  1 -1 -1 -1  1 -1 -1  1\n",
      "  1  1  1  1  1 -1  1  1  1  1  1 -1  1  1  1  1  1 -1  1  1  1  1  1 -1\n",
      "  1  1  1 -1 -1 -1 -1 -1  1  1  1  1 -1  1 -1  1 -1  1  1  1  1 -1  1  1\n",
      "  1  1  1  1  1  1  1 -1  1  1 -1  1  1  1 -1 -1 -1 -1 -1 -1  1  1 -1  1\n",
      "  1  1  1 -1  1 -1 -1 -1 -1  1  1 -1 -1 -1  1 -1 -1  1  1 -1 -1  1 -1 -1\n",
      " -1 -1 -1 -1 -1 -1  1 -1  1 -1 -1  1  1 -1 -1 -1 -1  1 -1 -1 -1  1  1  1\n",
      " -1  1 -1  1 -1  1 -1 -1 -1  1 -1 -1  1 -1 -1  1 -1  1  1  1  1  1 -1 -1\n",
      " -1 -1 -1 -1  1 -1 -1 -1  1  1 -1 -1 -1  1 -1 -1  1  1  1 -1  1  1 -1 -1\n",
      " -1 -1  1  1 -1 -1  1  1 -1  1  1  1 -1  1 -1 -1 -1 -1 -1 -1  1  1 -1 -1\n",
      "  1  1 -1 -1 -1  1  1 -1 -1 -1  1 -1  1  1 -1  1 -1  1  1  1  1  1 -1  1\n",
      " -1  1  1  1 -1 -1  1 -1  1  1 -1  1  1 -1  1  1  1  1  1 -1 -1 -1 -1  1\n",
      "  1  1 -1 -1  1 -1  1  1  1 -1 -1 -1  1  1 -1  1  1 -1  1  1 -1 -1 -1  1\n",
      "  1 -1  1 -1  1  1 -1 -1 -1 -1  1 -1 -1 -1 -1  1 -1  1  1 -1  1  1  1  1\n",
      "  1 -1 -1  1  1 -1  1 -1 -1  1  1 -1 -1 -1 -1  1 -1 -1  1  1 -1 -1 -1 -1\n",
      " -1  1 -1 -1  1  1 -1 -1 -1  1  1 -1 -1 -1  1 -1 -1 -1  1 -1  1 -1 -1  1\n",
      " -1 -1  1  1  1 -1 -1 -1 -1 -1  1  1  1 -1  1  1 -1 -1  1  1  1 -1  1  1\n",
      "  1  1  1  1 -1  1 -1  1  1  1 -1 -1  1  1  1  1  1 -1  1 -1  1  1  1  1\n",
      "  1  1  1 -1  1 -1 -1 -1 -1  1  1  1  1  1  1 -1 -1  1  1  1 -1 -1 -1 -1\n",
      "  1 -1  1 -1 -1  1 -1 -1  1  1  1 -1  1  1  1  1 -1  1  1  1 -1  1 -1 -1\n",
      " -1  1 -1  1  1  1  1  1 -1 -1  1 -1  1  1 -1 -1 -1  1  1 -1 -1  1  1 -1\n",
      " -1  1 -1 -1 -1 -1  1  1 -1  1 -1  1 -1  1 -1  1  1  1  1 -1  1 -1 -1  1\n",
      "  1  1 -1  1 -1  1  1 -1  1  1 -1 -1  1 -1  1  1  1 -1  1  1  1 -1 -1  1\n",
      "  1  1  1 -1 -1  1 -1  1  1  1 -1 -1 -1 -1 -1  1  1 -1 -1  1 -1 -1 -1 -1\n",
      " -1 -1 -1  1  1 -1 -1 -1 -1 -1  1  1  1  1 -1  1 -1 -1 -1 -1 -1  1  1 -1\n",
      "  1 -1  1  1  1 -1 -1 -1 -1 -1  1  1 -1 -1 -1 -1 -1 -1  1 -1 -1  1 -1  1\n",
      " -1 -1  1 -1  1 -1 -1 -1 -1  1  1 -1 -1  1 -1  1 -1 -1 -1  1  1 -1 -1 -1\n",
      " -1  1 -1  1 -1 -1  1 -1 -1  1  1  1 -1  1  1 -1  1  1  1  1  1 -1 -1 -1\n",
      "  1 -1 -1  1  1  1 -1 -1  1  1 -1 -1 -1  1  1  1 -1  1  1 -1  1  1  1  1\n",
      " -1  1  1 -1 -1  1  1 -1 -1  1  1  1  1 -1  1 -1 -1  1  1 -1 -1 -1  1  1\n",
      " -1  1  1 -1  1 -1  1 -1  1  1 -1 -1 -1 -1  1  1 -1 -1  1 -1 -1 -1  1 -1\n",
      "  1  1  1  1  1  1 -1 -1 -1  1 -1 -1 -1  1  1  1 -1 -1 -1  1  1  1  1  1\n",
      " -1 -1  1 -1  1  1  1 -1 -1  1  1 -1 -1  1  1 -1 -1  1  1 -1  1 -1  1  1\n",
      "  1 -1 -1 -1 -1 -1  1  1  1 -1  1 -1  1  1 -1 -1 -1  1  1  1  1 -1  1 -1\n",
      " -1 -1 -1 -1 -1  1  1 -1 -1  1  1  1 -1  1 -1  1  1  1 -1  1  1 -1 -1 -1\n",
      "  1 -1  1  1 -1  1  1 -1 -1 -1 -1  1 -1  1  1 -1 -1 -1 -1 -1 -1 -1  1 -1\n",
      " -1  1 -1 -1 -1  1  1 -1  1  1 -1 -1  1 -1 -1  1]\n"
     ]
    }
   ],
   "source": [
    "print(y)\n",
    "\n",
    "y = np.where(y ==0, -1, 1)\n",
    "\n",
    "print(y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "x = [ 1.9738175  -7.50934555] , y = 1\n",
      "x = [-1.19505856 -5.8378053 ] , y = -1\n",
      "x = [  0.37631303 -13.57342824] , y = -1\n",
      "x = [-0.58648176 -0.37741265] , y = 1\n",
      "x = [ 5.5126432  -1.76458642] , y = 1\n"
     ]
    }
   ],
   "source": [
    "for i in range(5):\n",
    "    print('x =', X[i, ], ', y =', y[i])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([ 1, -1, -1,  1,  1, -1,  1, -1,  1,  1])"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Just to show you what is inside it. We print the first 10 elements of this numpy array. \n",
    "y[:10]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# scatter plot, data points annotated by different colors\n",
    "df = pd.DataFrame(dict(feature_1=X[:,0], feature_2=X[:,1], label=y))\n",
    "\n",
    "\n",
    "cluster_name = set(y)\n",
    "colors = dict(zip(cluster_name, cm.rainbow(np.linspace(0, 1, len(cluster_name)))))\n",
    "fig, ax = pyplot.subplots()\n",
    "\n",
    "grouped = df.groupby('label')\n",
    "\n",
    "for key, group in grouped:\n",
    "    group.plot(ax=ax, kind='scatter', x='feature_1', y='feature_2', label=key, color=colors[key].reshape(1,-1))\n",
    "    \n",
    "pyplot.title('Original 2D Data from {} Clusters'.format(n_components))\n",
    "pyplot.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "\n",
    "# Create Test/Train\n",
    "from sklearn.model_selection import train_test_split\n",
    "\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=123)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Support Vector Machine Using Scikit-learn \n",
    "\n",
    "\n",
    "## SVC\n",
    "\n",
    "\n",
    "# Read for more details the following important documentations\n",
    "\n",
    "\n",
    "https://scikit-learn.org/stable/modules/svm.html#svm-classification\n",
    "\n",
    "\n",
    "\n",
    "https://scikit-learn.org/stable/modules/generated/sklearn.svm.SVC.html#sklearn.svm.SVC"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy Score: 0.875\n"
     ]
    }
   ],
   "source": [
    "\n",
    "from sklearn import svm\n",
    "from sklearn.metrics import accuracy_score\n",
    "\n",
    "\n",
    "model = svm.SVC()\n",
    "model.fit(X_train, y_train)\n",
    "\n",
    "\n",
    "\n",
    "y_pred = model.predict(X_test)\n",
    "\n",
    "print(\"Accuracy Score:\", accuracy_score(y_test, y_pred))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "          -1       0.87      0.89      0.88       104\n",
      "           1       0.88      0.85      0.87        96\n",
      "\n",
      "    accuracy                           0.88       200\n",
      "   macro avg       0.88      0.87      0.87       200\n",
      "weighted avg       0.88      0.88      0.87       200\n",
      "\n"
     ]
    }
   ],
   "source": [
    "from sklearn.metrics import classification_report\n",
    "print(classification_report(y_test,y_pred))\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Linear SVC \n",
    "\n",
    "## sklearn.svm.LinearSVC\n",
    "\n",
    "https://scikit-learn.org/stable/modules/generated/sklearn.svm.LinearSVC.html\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['C', '__class__', '__delattr__', '__dict__', '__dir__', '__doc__', '__eq__', '__format__', '__ge__', '__getattribute__', '__getstate__', '__gt__', '__hash__', '__init__', '__init_subclass__', '__le__', '__lt__', '__module__', '__ne__', '__new__', '__reduce__', '__reduce_ex__', '__repr__', '__setattr__', '__setstate__', '__sizeof__', '__str__', '__subclasshook__', '__weakref__', '_check_n_features', '_estimator_type', '_get_param_names', '_get_tags', '_more_tags', '_predict_proba_lr', '_repr_html_', '_repr_html_inner', '_repr_mimebundle_', '_validate_data', 'class_weight', 'classes_', 'coef_', 'decision_function', 'densify', 'dual', 'fit', 'fit_intercept', 'get_params', 'intercept_', 'intercept_scaling', 'loss', 'max_iter', 'multi_class', 'n_features_in_', 'n_iter_', 'penalty', 'predict', 'random_state', 'score', 'set_params', 'sparsify', 'tol', 'verbose']\n",
      "[[0.05907893 0.06076751]]\n",
      "False\n",
      "Accuracy Score: 0.73\n"
     ]
    }
   ],
   "source": [
    "model = svm.LinearSVC(C=.01, max_iter=100, fit_intercept=False)\n",
    "model.fit(X_train, y_train)\n",
    "\n",
    "print(dir(model))\n",
    "\n",
    "print(model.coef_)\n",
    "print(model.fit_intercept)\n",
    "\n",
    "y_pred = model.predict(X_test)\n",
    "\n",
    "print(\"Accuracy Score:\", accuracy_score(y_test, y_pred))\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# SVM in Python without libraries (no Scikit-learn) "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Cost Function \n",
    "\n",
    "Hinge Loss function\n",
    "\n",
    "\n",
    "$$ \n",
    "cost = \\frac{ \\lambda }{ 2 } ||w||^2 + \\frac{1}{N} \\sum_i^n max(0, 1 - y_i (w \\cdot x_i + b )) \\\\\n",
    "\\text{     Regularization parameter     }  \\lambda = \\frac{1}{c} \n",
    "$$"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {},
   "outputs": [],
   "source": [
    "def compute_cost(X, y, W, regularization_factor):\n",
    "    '''This function calculate the hinge loss. Primal Problem in SVM'''\n",
    "    n = X.shape[0]\n",
    "    distances = 1 - y * (np.dot(X, W))\n",
    "    \n",
    "    # This is our max(0, distance). \n",
    "    distances[distances < 0] = 0 \n",
    "    \n",
    "    hinge_loss = regularization_factor * (np.sum(distances) / n)\n",
    "    # This divide by 2 is not important. You can skip doing it \n",
    "    # because we want only to check if this cost is going down or not. \n",
    "    return (1 / 2 * np.dot(W, W) + hinge_loss)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Gradient \n",
    "$$gradient(w)=\\frac{1}{N}\\sum_i^n\n",
    "\\begin{cases}\n",
    "w &  \\text{ if } max(0, y_i (w \\cdot x_i)) = 0 \\\\\n",
    "w - c y_i x_i &  \\text{otherwise}\n",
    "\\end{cases}\n",
    "$$\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "metadata": {},
   "outputs": [],
   "source": [
    "def calculate_gradient(X, y, W, regularization_factor):\n",
    "      \n",
    "    if type(y) == np.float64:\n",
    "        y = np.array([y])\n",
    "        X = np.array([X])\n",
    "        \n",
    "    distance = 1 - (y * np.dot(X, W))\n",
    "    \n",
    "    dw = np.zeros(len(W))\n",
    "    \n",
    "    for ind, d in enumerate(distance):\n",
    "        \n",
    "        if (d < 0):\n",
    "            di = W\n",
    "        else:\n",
    "            di = W - (regularization_factor * y[ind] * X[ind])\n",
    "            \n",
    "            \n",
    "        dw += di\n",
    "    \n",
    "    dw = dw/len(y)  # average\n",
    "    \n",
    "    return dw"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 66,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 0 Cost is: 0.01 weights [0. 0.]\n",
      "Epoch 1 Cost is: 0.009981766434693546 weights [0.00013039 0.00040774]\n",
      "Epoch 2 Cost is: 0.009963895717336696 weights [0.00025949 0.0008114 ]\n",
      "Epoch 3 Cost is: 0.009946380627255244 weights [0.00038729 0.00121102]\n",
      "Epoch 4 Cost is: 0.009929214087466411 weights [0.00051381 0.00160664]\n",
      "Epoch 5 Cost is: 0.009912389161819378 weights [0.00063907 0.00199831]\n",
      "Epoch 6 Cost is: 0.00989589905219272 weights [0.00076307 0.00238607]\n",
      "Epoch 7 Cost is: 0.009879737095747633 weights [0.00088583 0.00276994]\n",
      "Epoch 8 Cost is: 0.009863896762235804 weights [0.00100737 0.00314998]\n",
      "Epoch 9 Cost is: 0.009848371651360861 weights [0.00112769 0.00352622]\n",
      "Epoch 10 Cost is: 0.009833155490192327 weights [0.00124681 0.00389869]\n",
      "Epoch 11 Cost is: 0.009818242130631049 weights [0.00136474 0.00426744]\n",
      "Epoch 12 Cost is: 0.00980362554692504 weights [0.00148148 0.0046325 ]\n",
      "Epoch 13 Cost is: 0.009789299833234778 weights [0.00159706 0.00499391]\n",
      "Epoch 14 Cost is: 0.009775259201246955 weights [0.00171149 0.00535171]\n",
      "Epoch 15 Cost is: 0.00976149797783569 weights [0.00182477 0.00570593]\n",
      "Epoch 16 Cost is: 0.009748010602770306 weights [0.00193692 0.00605661]\n",
      "Epoch 17 Cost is: 0.009734791626468725 weights [0.00204794 0.00640378]\n",
      "Epoch 18 Cost is: 0.009721835707795546 weights [0.00215786 0.00674748]\n",
      "Epoch 19 Cost is: 0.009709137611903961 weights [0.00226667 0.00708774]\n",
      "Epoch 20 Cost is: 0.009696692208120623 weights [0.0023744 0.0074246]\n",
      "Epoch 21 Cost is: 0.009684494467872569 weights [0.00248105 0.00775809]\n",
      "Epoch 22 Cost is: 0.009672539462655455 weights [0.00258664 0.00808824]\n",
      "Epoch 23 Cost is: 0.00966082236204216 weights [0.00269117 0.0084151 ]\n",
      "Epoch 24 Cost is: 0.009649338431731068 weights [0.00279465 0.00873868]\n",
      "Epoch 25 Cost is: 0.009638083031633169 weights [0.0028971  0.00905903]\n",
      "Epoch 26 Cost is: 0.009627051613997217 weights [0.00299852 0.00937618]\n",
      "Epoch 27 Cost is: 0.009616239721572222 weights [0.00309893 0.00969016]\n",
      "Epoch 28 Cost is: 0.009605642985806481 weights [0.00319834 0.01000099]\n",
      "Epoch 29 Cost is: 0.009595257125082482 weights [0.00329675 0.01030872]\n",
      "Epoch 30 Cost is: 0.009585077942986887 weights [0.00339418 0.01061337]\n",
      "Epoch 31 Cost is: 0.009575101326614996 weights [0.00349063 0.01091497]\n",
      "Epoch 32 Cost is: 0.009565323244908904 weights [0.00358612 0.01121356]\n",
      "Epoch 33 Cost is: 0.009555739747028768 weights [0.00368065 0.01150916]\n",
      "Epoch 34 Cost is: 0.009546346960756442 weights [0.00377424 0.0118018 ]\n",
      "Epoch 35 Cost is: 0.009537141090930939 weights [0.00386689 0.01209152]\n",
      "Epoch 36 Cost is: 0.00952811841791496 weights [0.00395862 0.01237834]\n",
      "Epoch 37 Cost is: 0.009519275296092 weights [0.00404943 0.01266229]\n",
      "Epoch 38 Cost is: 0.009510608152393319 weights [0.00413933 0.01294341]\n",
      "Epoch 39 Cost is: 0.009502113484854238 weights [0.00422833 0.01322171]\n",
      "Epoch 40 Cost is: 0.009493787861199188 weights [0.00431644 0.01349723]\n",
      "Epoch 41 Cost is: 0.009485627917454872 weights [0.00440367 0.01376999]\n",
      "Epoch 42 Cost is: 0.009477630356591068 weights [0.00449003 0.01404003]\n",
      "Epoch 43 Cost is: 0.009469791947188454 weights [0.00457553 0.01430737]\n",
      "Epoch 44 Cost is: 0.009462109522132953 weights [0.00466017 0.01457203]\n",
      "Epoch 45 Cost is: 0.009454579977336055 weights [0.00474396 0.01483405]\n",
      "Epoch 46 Cost is: 0.009447200270480615 weights [0.00482691 0.01509344]\n",
      "Epoch 47 Cost is: 0.0094399674197916 weights [0.00490904 0.01535024]\n",
      "Epoch 48 Cost is: 0.009432878502831294 weights [0.00499035 0.01560448]\n",
      "Epoch 49 Cost is: 0.0094259306553185 weights [0.00507084 0.01585617]\n",
      "Epoch 50 Cost is: 0.009419121069971212 weights [0.00515052 0.01610534]\n",
      "Epoch 51 Cost is: 0.009412446995372333 weights [0.00522941 0.01635203]\n",
      "Epoch 52 Cost is: 0.009405905734857972 weights [0.00530751 0.01659624]\n",
      "Epoch 53 Cost is: 0.009399494645427846 weights [0.00538483 0.01683802]\n",
      "Epoch 54 Cost is: 0.00939321113667738 weights [0.00546138 0.01707737]\n",
      "Epoch 55 Cost is: 0.009387052669751048 weights [0.00553716 0.01731434]\n",
      "Epoch 56 Cost is: 0.009381016756316552 weights [0.00561219 0.01754893]\n",
      "Epoch 57 Cost is: 0.009375100957559399 weights [0.00568646 0.01778118]\n",
      "Epoch 58 Cost is: 0.009369302883197515 weights [0.00575999 0.0180111 ]\n",
      "Epoch 59 Cost is: 0.009363620190515435 weights [0.00583278 0.01823873]\n",
      "Epoch 60 Cost is: 0.009358050583417724 weights [0.00590485 0.01846408]\n",
      "Epoch 61 Cost is: 0.00935259181150126 weights [0.0059762  0.01868717]\n",
      "Epoch 62 Cost is: 0.009347241669145932 weights [0.00604683 0.01890804]\n",
      "Epoch 63 Cost is: 0.009341997994623479 weights [0.00611676 0.01912669]\n",
      "Epoch 64 Cost is: 0.009336858669224018 weights [0.00618598 0.01934316]\n",
      "Epoch 65 Cost is: 0.00933182161640001 weights [0.00625452 0.01955747]\n",
      "Epoch 66 Cost is: 0.009326884800927197 weights [0.00632237 0.01976963]\n",
      "Epoch 67 Cost is: 0.009322046228082293 weights [0.00638954 0.01997967]\n",
      "Epoch 68 Cost is: 0.009317303942837005 weights [0.00645604 0.02018761]\n",
      "Epoch 69 Cost is: 0.009312656029068096 weights [0.00652188 0.02039347]\n",
      "Epoch 70 Cost is: 0.009308100608783187 weights [0.00658705 0.02059727]\n",
      "Epoch 71 Cost is: 0.009303635841361952 weights [0.00665158 0.02079904]\n",
      "Epoch 72 Cost is: 0.009299259922812398 weights [0.00671546 0.02099878]\n",
      "Epoch 73 Cost is: 0.00929497108504198 weights [0.0067787  0.02119653]\n",
      "Epoch 74 Cost is: 0.00929076759514319 weights [0.0068413 0.0213923]\n",
      "Epoch 75 Cost is: 0.00928664775469339 weights [0.00690329 0.02158612]\n",
      "Epoch 76 Cost is: 0.00928260989906854 weights [0.00696465 0.02177799]\n",
      "Epoch 77 Cost is: 0.009278652396770625 weights [0.0070254  0.02196795]\n",
      "Epoch 78 Cost is: 0.009274773648768437 weights [0.00708554 0.022156  ]\n",
      "Epoch 79 Cost is: 0.009270972087851496 weights [0.00714508 0.02234218]\n",
      "Epoch 80 Cost is: 0.009267246177996794 weights [0.00720402 0.0225265 ]\n",
      "Epoch 81 Cost is: 0.00926359441374821 weights [0.00726238 0.02270897]\n",
      "Epoch 82 Cost is: 0.009260015319608166 weights [0.00732015 0.02288961]\n",
      "Epoch 83 Cost is: 0.009256507449441512 weights [0.00737734 0.02306845]\n",
      "Epoch 84 Cost is: 0.009253069385891176 weights [0.00743396 0.02324551]\n",
      "Epoch 85 Cost is: 0.00924969973980549 weights [0.00749002 0.02342079]\n",
      "Epoch 86 Cost is: 0.009246397149676906 weights [0.00754551 0.02359432]\n",
      "Epoch 87 Cost is: 0.009243160281091888 weights [0.00760045 0.02376611]\n",
      "Epoch 88 Cost is: 0.009239987826191705 weights [0.00765484 0.02393618]\n",
      "Epoch 89 Cost is: 0.009236878503144037 weights [0.00770869 0.02410456]\n",
      "Epoch 90 Cost is: 0.009233831055625022 weights [0.007762   0.02427125]\n",
      "Epoch 91 Cost is: 0.00923084425231163 weights [0.00781477 0.02443627]\n",
      "Epoch 92 Cost is: 0.009227916886384177 weights [0.00786702 0.02459965]\n",
      "Epoch 93 Cost is: 0.009225047775038681 weights [0.00791874 0.02476139]\n",
      "Epoch 94 Cost is: 0.00922223575900896 weights [0.00796995 0.02492151]\n",
      "Epoch 95 Cost is: 0.00921947970209823 weights [0.00802065 0.02508003]\n",
      "Epoch 96 Cost is: 0.009216778490720022 weights [0.00807084 0.02523697]\n",
      "Epoch 97 Cost is: 0.009214131033448243 weights [0.00812052 0.02539234]\n",
      "Epoch 98 Cost is: 0.009211536260576173 weights [0.00816971 0.02554615]\n",
      "Epoch 99 Cost is: 0.009208993123684252 weights [0.00821841 0.02569842]\n",
      "Epoch 100 Cost is: 0.009206500595216485 weights [0.00826662 0.02584918]\n",
      "Epoch 101 Cost is: 0.009204057668065225 weights [0.00831435 0.02599842]\n",
      "Epoch 102 Cost is: 0.009201663355164277 weights [0.0083616  0.02614617]\n",
      "Epoch 103 Cost is: 0.009199316689090055 weights [0.00840838 0.02629245]\n",
      "Epoch 104 Cost is: 0.009197016721670712 weights [0.00845469 0.02643726]\n",
      "Epoch 105 Cost is: 0.00919476252360301 weights [0.00850054 0.02658062]\n",
      "Epoch 106 Cost is: 0.009192553184076859 weights [0.00854593 0.02672255]\n",
      "Epoch 107 Cost is: 0.009190387810407277 weights [0.00859087 0.02686306]\n",
      "Epoch 108 Cost is: 0.009188265527673722 weights [0.00863535 0.02700217]\n",
      "Epoch 109 Cost is: 0.009186185478366563 weights [0.00867939 0.02713989]\n",
      "Epoch 110 Cost is: 0.009184146822040617 weights [0.00872299 0.02727622]\n",
      "Epoch 111 Cost is: 0.009182148734975557 weights [0.00876616 0.0274112 ]\n",
      "Epoch 112 Cost is: 0.009180190409843091 weights [0.00880889 0.02754482]\n",
      "Epoch 113 Cost is: 0.009178271055380762 weights [0.0088512  0.02767711]\n",
      "Epoch 114 Cost is: 0.009176389896072232 weights [0.00889308 0.02780808]\n",
      "Epoch 115 Cost is: 0.009174546171833945 weights [0.00893455 0.02793773]\n",
      "Epoch 116 Cost is: 0.009172739137707997 weights [0.0089756  0.02806609]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 117 Cost is: 0.009170968063561156 weights [0.00901623 0.02819317]\n",
      "Epoch 118 Cost is: 0.009169232233789838 weights [0.00905647 0.02831897]\n",
      "Epoch 119 Cost is: 0.009167530947030968 weights [0.0090963  0.02844352]\n",
      "Epoch 120 Cost is: 0.009165863515878602 weights [0.00913573 0.02856682]\n",
      "Epoch 121 Cost is: 0.009164229266606164 weights [0.00917477 0.02868889]\n",
      "Epoch 122 Cost is: 0.00916262753889425 weights [0.00921341 0.02880973]\n",
      "Epoch 123 Cost is: 0.0091610576855638 weights [0.00925168 0.02892937]\n",
      "Epoch 124 Cost is: 0.009159519072314631 weights [0.00928955 0.02904782]\n",
      "Epoch 125 Cost is: 0.009158011077469117 weights [0.00932705 0.02916508]\n",
      "Epoch 126 Cost is: 0.00915653309172103 weights [0.00936418 0.02928116]\n",
      "Epoch 127 Cost is: 0.009155084517889331 weights [0.00940093 0.02939609]\n",
      "Epoch 128 Cost is: 0.009153664770676879 weights [0.00943732 0.02950986]\n",
      "Epoch 129 Cost is: 0.00915227327643396 weights [0.00947334 0.0296225 ]\n",
      "Epoch 130 Cost is: 0.009150909472926473 weights [0.009509   0.02973401]\n",
      "Epoch 131 Cost is: 0.009149572809108783 weights [0.0095443  0.02984441]\n",
      "Epoch 132 Cost is: 0.009148262744901067 weights [0.00957926 0.0299537 ]\n",
      "Epoch 133 Cost is: 0.00914697875097108 weights [0.00961386 0.0300619 ]\n",
      "Epoch 134 Cost is: 0.009145720308520306 weights [0.00964812 0.03016902]\n",
      "Epoch 135 Cost is: 0.009144486909074302 weights [0.00968203 0.03027506]\n",
      "Epoch 136 Cost is: 0.009143278054277271 weights [0.0097156  0.03038005]\n",
      "Epoch 137 Cost is: 0.009142093255690702 weights [0.00974884 0.03048398]\n",
      "Epoch 138 Cost is: 0.009140932034596004 weights [0.00978175 0.03058688]\n",
      "Epoch 139 Cost is: 0.00913979392180109 weights [0.00981433 0.03068875]\n",
      "Epoch 140 Cost is: 0.009138678457450799 weights [0.00984658 0.0307896 ]\n",
      "Epoch 141 Cost is: 0.009137585190841076 weights [0.00987851 0.03088944]\n",
      "Epoch 142 Cost is: 0.009136513680236887 weights [0.00991012 0.03098828]\n",
      "Epoch 143 Cost is: 0.009135463492693719 weights [0.00994141 0.03108613]\n",
      "Epoch 144 Cost is: 0.009134434203882663 weights [0.00997239 0.03118301]\n",
      "Epoch 145 Cost is: 0.009133425397918947 weights [0.01000306 0.03127892]\n",
      "Epoch 146 Cost is: 0.009132436667193908 weights [0.01003343 0.03137386]\n",
      "Epoch 147 Cost is: 0.009131467612210299 weights [0.01006349 0.03146786]\n",
      "Epoch 148 Cost is: 0.009130517841420861 weights [0.01009325 0.03156092]\n",
      "Epoch 149 Cost is: 0.009129586971070133 weights [0.01012271 0.03165305]\n",
      "Epoch 150 Cost is: 0.009128674625039388 weights [0.01015188 0.03174425]\n",
      "Epoch 151 Cost is: 0.00912778043469465 weights [0.01018076 0.03183455]\n",
      "Epoch 152 Cost is: 0.009126904038737775 weights [0.01020934 0.03192394]\n",
      "Epoch 153 Cost is: 0.009126045083060442 weights [0.01023764 0.03201243]\n",
      "Epoch 154 Cost is: 0.009125203220601086 weights [0.01026566 0.03210005]\n",
      "Epoch 155 Cost is: 0.009124378111204675 weights [0.0102934  0.03218678]\n",
      "Epoch 156 Cost is: 0.00912356942148525 weights [0.01032086 0.03227265]\n",
      "Epoch 157 Cost is: 0.009122776824691241 weights [0.01034805 0.03235766]\n",
      "Epoch 158 Cost is: 0.009122000000573434 weights [0.01037496 0.03244182]\n",
      "Epoch 159 Cost is: 0.00912123863525557 weights [0.01040161 0.03252514]\n",
      "Epoch 160 Cost is: 0.009120492421107533 weights [0.01042799 0.03260762]\n",
      "Epoch 161 Cost is: 0.009119761056621042 weights [0.0104541  0.03268929]\n",
      "Epoch 162 Cost is: 0.00911904424628783 weights [0.01047996 0.03277013]\n",
      "Epoch 163 Cost is: 0.009118341700480252 weights [0.01050555 0.03285016]\n",
      "Epoch 164 Cost is: 0.009117653135334244 weights [0.01053089 0.0329294 ]\n",
      "Epoch 165 Cost is: 0.009116978272634638 weights [0.01055598 0.03300784]\n",
      "Epoch 166 Cost is: 0.009116316839702758 weights [0.01058081 0.0330855 ]\n",
      "Epoch 167 Cost is: 0.009115668569286223 weights [0.0106054  0.03316238]\n",
      "Epoch 168 Cost is: 0.009115033199450973 weights [0.01062974 0.03323849]\n",
      "Epoch 169 Cost is: 0.00911441047347545 weights [0.01065384 0.03331385]\n",
      "Epoch 170 Cost is: 0.009113800139746834 weights [0.0106777  0.03338844]\n",
      "Epoch 171 Cost is: 0.009113201951659421 weights [0.01070131 0.0334623 ]\n",
      "Epoch 172 Cost is: 0.009112615667514945 weights [0.01072469 0.03353541]\n",
      "Epoch 173 Cost is: 0.009112041050424947 weights [0.01074784 0.03360779]\n",
      "Epoch 174 Cost is: 0.00911147786821504 weights [0.01077076 0.03367945]\n",
      "Epoch 175 Cost is: 0.00911092589333111 weights [0.01079345 0.03375039]\n",
      "Epoch 176 Cost is: 0.009110384902747366 weights [0.01081591 0.03382062]\n",
      "Epoch 177 Cost is: 0.009109854677876244 weights [0.01083814 0.03389015]\n",
      "Epoch 178 Cost is: 0.009109335004480053 weights [0.01086016 0.03395899]\n",
      "Epoch 179 Cost is: 0.00910882567258445 weights [0.01088195 0.03402714]\n",
      "Epoch 180 Cost is: 0.009108326476393567 weights [0.01090353 0.0340946 ]\n",
      "Epoch 181 Cost is: 0.009107837214206882 weights [0.01092489 0.03416139]\n",
      "Epoch 182 Cost is: 0.009107357688337717 weights [0.01094603 0.03422751]\n",
      "Epoch 183 Cost is: 0.009106887705033342 weights [0.01096697 0.03429298]\n",
      "Epoch 184 Cost is: 0.009106427074396727 weights [0.01098769 0.03435778]\n",
      "Epoch 185 Cost is: 0.00910597561030978 weights [0.01100821 0.03442194]\n",
      "Epoch 186 Cost is: 0.009105533130358164 weights [0.01102852 0.03448546]\n",
      "Epoch 187 Cost is: 0.009105099455757586 weights [0.01104863 0.03454834]\n",
      "Epoch 188 Cost is: 0.009104674411281558 weights [0.01106854 0.03461059]\n",
      "Epoch 189 Cost is: 0.009104257825190603 weights [0.01108825 0.03467222]\n",
      "Epoch 190 Cost is: 0.009103849529162857 weights [0.01110776 0.03473324]\n",
      "Epoch 191 Cost is: 0.009103449358226063 weights [0.01112708 0.03479364]\n",
      "Epoch 192 Cost is: 0.009103057150690914 weights [0.0111462  0.03485344]\n",
      "Epoch 193 Cost is: 0.009102672748085714 weights [0.01116514 0.03491264]\n",
      "Epoch 194 Cost is: 0.009102295995092354 weights [0.01118388 0.03497125]\n",
      "Epoch 195 Cost is: 0.009101926739483568 weights [0.01120244 0.03502928]\n",
      "Epoch 196 Cost is: 0.00910156483206139 weights [0.01122081 0.03508672]\n",
      "Epoch 197 Cost is: 0.009101210126596917 weights [0.011239   0.03514359]\n",
      "Epoch 198 Cost is: 0.009100862479771188 weights [0.011257   0.03519989]\n",
      "Epoch 199 Cost is: 0.009100521751117288 weights [0.01127483 0.03525563]\n"
     ]
    }
   ],
   "source": [
    "weights = np.zeros(2)\n",
    "\n",
    "\n",
    "# Now we optimize it using Gradient Descent. \n",
    "num_iterations = 200\n",
    "learnin_rate = 0.01\n",
    "\n",
    "regularization = 0.01\n",
    "\n",
    "cost_list = []\n",
    "\n",
    "for i in range(0, num_iterations):\n",
    "    \n",
    "    cost = compute_cost(X_train, y_train, weights, regularization)\n",
    "    \n",
    "    print(\"Epoch\", i, \"Cost is:\", cost, \"weights\", weights)\n",
    "    \n",
    "    cost_list.append(cost)\n",
    "    \n",
    "    grad = calculate_gradient(X_train, y_train, weights, regularization)\n",
    "    \n",
    "    weights = weights - learnin_rate * grad"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 67,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(np.arange(num_iterations), cost_list)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 68,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([-1, -1, -1, -1, -1, -1,  1, -1,  1, -1, -1, -1,  1, -1, -1, -1, -1,\n",
       "        1, -1, -1, -1, -1,  1, -1, -1, -1, -1, -1, -1, -1,  1,  1, -1, -1,\n",
       "       -1, -1, -1, -1, -1,  1,  1, -1,  1, -1, -1, -1, -1, -1, -1, -1, -1,\n",
       "       -1,  1, -1, -1, -1,  1, -1,  1, -1, -1, -1, -1,  1, -1, -1, -1, -1,\n",
       "       -1, -1, -1, -1, -1, -1, -1,  1, -1, -1,  1,  1, -1, -1, -1, -1, -1,\n",
       "       -1,  1, -1, -1, -1, -1, -1,  1, -1, -1, -1,  1, -1,  1, -1, -1, -1,\n",
       "       -1,  1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,  1, -1, -1,\n",
       "       -1, -1,  1, -1, -1, -1, -1, -1, -1,  1, -1,  1,  1, -1, -1, -1, -1,\n",
       "       -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,  1, -1,\n",
       "       -1, -1, -1, -1, -1, -1, -1, -1,  1, -1, -1, -1,  1, -1, -1, -1, -1,\n",
       "       -1, -1, -1,  1, -1, -1, -1, -1, -1, -1,  1, -1,  1, -1,  1,  1, -1,\n",
       "       -1, -1, -1, -1, -1,  1, -1, -1, -1, -1,  1,  1, -1])"
      ]
     },
     "execution_count": 68,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Prediction calculculation \n",
    "# print(np.dot(X_test, weights))\n",
    "y_predict_2 = np.where(np.dot(X_test, weights)<0, -1, 1)\n",
    "\n",
    "y_predict_2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 69,
   "metadata": {},
   "outputs": [],
   "source": [
    "def predict_svm(weights, X):\n",
    "    \"\"\"\n",
    "    Predict the class between 0 and 1 using learned SVM parameters weights.\n",
    "    \"\"\"    \n",
    "    return np.where(np.dot(X, weights)<0, -1, 1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 70,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy Score: 0.71\n"
     ]
    }
   ],
   "source": [
    "print(\"Accuracy Score:\", accuracy_score(y_test, predict_svm(weights, X_test)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.9.7"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}