stock_price_predictors/deep_learning_models/lstm_stock_predictor.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "source": [
    "# LSTM Stock Predictor Using Closing Prices\n",
    "\n",
    "In this notebook, we will build and train a custom LSTM RNN that uses a 10 day window of chosen stock closing prices to predict the 11th day closing price. \n",
    "\n",
    "We will need to:\n",
    "\n",
    "1. Prepare the data for training and testing\n",
    "2. Build and train a custom LSTM RNN\n",
    "3. Evaluate the performance of the model"
   ],
   "metadata": {}
  },
  {
   "cell_type": "markdown",
   "source": [
    "## Data Preparation\n",
    "\n",
    "In this section, we will need to prepare the training and testing data for the model. The model will use a rolling 10 day window to predict the 11th day closing price.\n",
    "\n",
    "We will need to:\n",
    "1. Use the `window_data` function to generate the X and y values for the model.\n",
    "2. Split the data into 70% training and 30% testing\n",
    "3. Apply the MinMaxScaler to the X and y values\n",
    "4. Reshape the X_train and X_test data for the model. Note: The required input format for the LSTM is:\n",
    "\n",
    "```python\n",
    "reshape((X_train.shape[0], X_train.shape[1], 1))\n",
    "```"
   ],
   "metadata": {}
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "%matplotlib inline"
   ],
   "outputs": [
    {
     "output_type": "stream",
     "name": "stderr",
     "text": [
      "\n",
      "Bad key \"text.kerning_factor\" on line 4 in\n",
      "/Users/draganbogatic/opt/anaconda3/envs/pyvizenv/lib/python3.7/site-packages/matplotlib/mpl-data/stylelib/_classic_test_patch.mplstyle.\n",
      "You probably need to get an updated matplotlibrc file from\n",
      "https://github.com/matplotlib/matplotlib/blob/master/matplotlibrc.template\n",
      "or from the matplotlib source distribution\n"
     ]
    }
   ],
   "metadata": {}
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "source": [
    "# Set the random seed for reproducibility\n",
    "# Note: It is a good practice to run multiple experiments to evaluate model\n",
    "#from numpy.random import seed\n",
    "#seed(1)\n",
    "#from tensorflow import random\n",
    "#random.set_seed(2)"
   ],
   "outputs": [],
   "metadata": {}
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "source": [
    "# Load the historical closing prices for the stock\n",
    "stock_df = pd.read_csv(\"../Resources/stock_historical_prices.csv\", index_col=\"Date\", infer_datetime_format=True, parse_dates=True)\n",
    "stock_df = stock_df.sort_index()\n",
    "\n",
    "stock_df.tail()"
   ],
   "outputs": [
    {
     "output_type": "execute_result",
     "data": {
      "text/plain": [
       "             Adj Close\n",
       "Date                  \n",
       "2021-07-13  279.209991\n",
       "2021-07-14  278.309998\n",
       "2021-07-15  279.720001\n",
       "2021-07-16  277.899994\n",
       "2021-07-19  272.660004"
      ],
      "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>Adj Close</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Date</th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2021-07-13</th>\n",
       "      <td>279.209991</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2021-07-14</th>\n",
       "      <td>278.309998</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2021-07-15</th>\n",
       "      <td>279.720001</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2021-07-16</th>\n",
       "      <td>277.899994</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2021-07-19</th>\n",
       "      <td>272.660004</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ]
     },
     "metadata": {},
     "execution_count": 7
    }
   ],
   "metadata": {}
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "source": [
    "# This function accepts the column number for the features (X) and the target (y)\n",
    "# It chunks the data up with a rolling window of Xt-n to predict Xt\n",
    "# It returns a numpy array of X any y\n",
    "def window_data(df, window, feature_col_number, target_col_number):\n",
    "    X = []\n",
    "    y = []\n",
    "    for i in range(len(df) - window - 1):\n",
    "        features = df.iloc[i:(i + window), feature_col_number]\n",
    "        target = df.iloc[(i + window), target_col_number]\n",
    "        X.append(features)\n",
    "        y.append(target)\n",
    "    return np.array(X), np.array(y).reshape(-1, 1)"
   ],
   "outputs": [],
   "metadata": {}
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "source": [
    "# Predict Closing Prices using a 10 day window of previous closing prices\n",
    "# Then, experiment with window sizes anywhere from 1 to 10 and see how the model performance changes\n",
    "window_size = 10\n",
    "\n",
    "# Column index 0 is the `close` column\n",
    "feature_column = 0\n",
    "target_column = 0\n",
    "X, y = window_data(stock_df, window_size, feature_column, target_column)\n",
    "\n",
    "print (f\"X sample values:\\n{X[:5]} \\n\")\n",
    "print (f\"y sample values:\\n{y[:5]}\")"
   ],
   "outputs": [
    {
     "output_type": "stream",
     "name": "stdout",
     "text": [
      "X sample values:\n",
      "[[143.55000305 143.41999817 142.80000305 141.6000061  141.94999695\n",
      "  142.57000732 142.32000732 141.36000061 140.47000122 138.38999939]\n",
      " [143.41999817 142.80000305 141.6000061  141.94999695 142.57000732\n",
      "  142.32000732 141.36000061 140.47000122 138.38999939 134.22000122]\n",
      " [142.80000305 141.6000061  141.94999695 142.57000732 142.32000732\n",
      "  141.36000061 140.47000122 138.38999939 134.22000122 129.57000732]\n",
      " [141.6000061  141.94999695 142.57000732 142.32000732 141.36000061\n",
      "  140.47000122 138.38999939 134.22000122 129.57000732 127.73999786]\n",
      " [141.94999695 142.57000732 142.32000732 141.36000061 140.47000122\n",
      "  138.38999939 134.22000122 129.57000732 127.73999786 132.61000061]] \n",
      "\n",
      "y sample values:\n",
      "[[134.22000122]\n",
      " [129.57000732]\n",
      " [127.73999786]\n",
      " [132.61000061]\n",
      " [136.22999573]]\n"
     ]
    }
   ],
   "metadata": {}
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "source": [
    "# Use 70% of the data for training and the remaineder for testing\n",
    "split = int(0.7 * len(X))\n",
    "X_train = X[: split]\n",
    "X_test = X[split:]\n",
    "y_train = y[: split]\n",
    "y_test = y[split:]"
   ],
   "outputs": [],
   "metadata": {}
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "source": [
    "from sklearn.preprocessing import MinMaxScaler\n",
    "# Use the MinMaxScaler to scale data between 0 and 1.\n",
    "scaler = MinMaxScaler()\n",
    "scaler.fit(X)\n",
    "X_train_scaler = scaler.transform(X_train)\n",
    "X_test_scaler = scaler.transform(X_test)\n",
    "scaler.fit(y)\n",
    "y_train = scaler.transform(y_train)\n",
    "y_test = scaler.transform(y_test)"
   ],
   "outputs": [],
   "metadata": {}
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "source": [
    "# Reshape the features for the model\n",
    "X_train = X_train_scaler.reshape((X_train_scaler.shape[0], X_train_scaler.shape[1], 1))\n",
    "X_test = X_test_scaler.reshape((X_test_scaler.shape[0], X_test_scaler.shape[1], 1))\n",
    "print (f\"X_train sample values:\\n{X_train[:5]} \\n\")\n",
    "print (f\"X_test sample values:\\n{X_test[:5]}\")"
   ],
   "outputs": [
    {
     "output_type": "stream",
     "name": "stdout",
     "text": [
      "X_train sample values:\n",
      "[[[0.11532755]\n",
      "  [0.11455551]\n",
      "  [0.11087361]\n",
      "  [0.10374731]\n",
      "  [0.10582576]\n",
      "  [0.10950776]\n",
      "  [0.10802311]\n",
      "  [0.10232202]\n",
      "  [0.09703667]\n",
      "  [0.08468438]]\n",
      "\n",
      " [[0.11455551]\n",
      "  [0.11087361]\n",
      "  [0.10374731]\n",
      "  [0.10582576]\n",
      "  [0.10950776]\n",
      "  [0.10802311]\n",
      "  [0.10232202]\n",
      "  [0.09703667]\n",
      "  [0.08468438]\n",
      "  [0.05992045]]\n",
      "\n",
      " [[0.11087361]\n",
      "  [0.10374731]\n",
      "  [0.10582576]\n",
      "  [0.10950776]\n",
      "  [0.10802311]\n",
      "  [0.10232202]\n",
      "  [0.09703667]\n",
      "  [0.08468438]\n",
      "  [0.05992045]\n",
      "  [0.03230602]]\n",
      "\n",
      " [[0.10374731]\n",
      "  [0.10582576]\n",
      "  [0.10950776]\n",
      "  [0.10802311]\n",
      "  [0.10232202]\n",
      "  [0.09703667]\n",
      "  [0.08468438]\n",
      "  [0.05992045]\n",
      "  [0.03230602]\n",
      "  [0.02143833]]\n",
      "\n",
      " [[0.10582576]\n",
      "  [0.10950776]\n",
      "  [0.10802311]\n",
      "  [0.10232202]\n",
      "  [0.09703667]\n",
      "  [0.08468438]\n",
      "  [0.05992045]\n",
      "  [0.03230602]\n",
      "  [0.02143833]\n",
      "  [0.05035931]]] \n",
      "\n",
      "X_test sample values:\n",
      "[[[0.49872325]\n",
      "  [0.49153751]\n",
      "  [0.49480375]\n",
      "  [0.49818879]\n",
      "  [0.49040925]\n",
      "  [0.46897092]\n",
      "  [0.48013541]\n",
      "  [0.4985451 ]\n",
      "  [0.4935567 ]\n",
      "  [0.47485007]]\n",
      "\n",
      " [[0.49153751]\n",
      "  [0.49480375]\n",
      "  [0.49818879]\n",
      "  [0.49040925]\n",
      "  [0.46897092]\n",
      "  [0.48013541]\n",
      "  [0.4985451 ]\n",
      "  [0.4935567 ]\n",
      "  [0.47485007]\n",
      "  [0.49029045]]\n",
      "\n",
      " [[0.49480375]\n",
      "  [0.49818879]\n",
      "  [0.49040925]\n",
      "  [0.46897092]\n",
      "  [0.48013541]\n",
      "  [0.4985451 ]\n",
      "  [0.4935567 ]\n",
      "  [0.47485007]\n",
      "  [0.49029045]\n",
      "  [0.4913    ]]\n",
      "\n",
      " [[0.49818879]\n",
      "  [0.49040925]\n",
      "  [0.46897092]\n",
      "  [0.48013541]\n",
      "  [0.4985451 ]\n",
      "  [0.4935567 ]\n",
      "  [0.47485007]\n",
      "  [0.49029045]\n",
      "  [0.4913    ]\n",
      "  [0.4985451 ]]\n",
      "\n",
      " [[0.49040925]\n",
      "  [0.46897092]\n",
      "  [0.48013541]\n",
      "  [0.4985451 ]\n",
      "  [0.4935567 ]\n",
      "  [0.47485007]\n",
      "  [0.49029045]\n",
      "  [0.4913    ]\n",
      "  [0.4985451 ]\n",
      "  [0.49753555]]]\n"
     ]
    }
   ],
   "metadata": {}
  },
  {
   "cell_type": "markdown",
   "source": [
    "---"
   ],
   "metadata": {}
  },
  {
   "cell_type": "markdown",
   "source": [
    "## Build and Train the LSTM RNN\n",
    "\n",
    "In this section, we will design a custom LSTM RNN and fit (train) it using the training data.\n",
    "\n",
    "Need to:\n",
    "1. Define the model architecture\n",
    "2. Compile the model\n",
    "3. Fit the model to the training data"
   ],
   "metadata": {}
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "source": [
    "from tensorflow.keras.models import Sequential\n",
    "from tensorflow.keras.layers import LSTM, Dense, Dropout"
   ],
   "outputs": [],
   "metadata": {}
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "source": [
    "# Build the LSTM model. \n",
    "# The return sequences need to be set to True if we are adding additional LSTM layers, but \n",
    "# We don't have to do this for the final layer. \n",
    "# Note: The dropouts help prevent overfitting\n",
    "# Note: The input shape is the number of time steps and the number of indicators\n",
    "# Note: Batching inputs has a different input shape of Samples/TimeSteps/Features\n",
    "\n",
    "model = Sequential()\n",
    "\n",
    "number_units = 30\n",
    "dropout_fraction = 0.2\n",
    "\n",
    "# Layer 1\n",
    "model.add(LSTM(\n",
    "    units=number_units,\n",
    "    return_sequences=True,\n",
    "    input_shape=(X_train.shape[1], 1))\n",
    "    )\n",
    "model.add(Dropout(dropout_fraction))\n",
    "# Layer 2\n",
    "model.add(LSTM(units=number_units, return_sequences=True))\n",
    "model.add(Dropout(dropout_fraction))\n",
    "# Layer 3\n",
    "model.add(LSTM(units=number_units))\n",
    "model.add(Dropout(dropout_fraction))\n",
    "# Output layer\n",
    "model.add(Dense(1))"
   ],
   "outputs": [],
   "metadata": {}
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "source": [
    "# Compile the model\n",
    "model.compile(optimizer=\"adam\", loss=\"mean_squared_error\")"
   ],
   "outputs": [],
   "metadata": {}
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "source": [
    "# Summarize the model\n",
    "model.summary()"
   ],
   "outputs": [
    {
     "output_type": "stream",
     "name": "stdout",
     "text": [
      "Model: \"sequential_1\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "lstm_3 (LSTM)                (None, 10, 30)            3840      \n",
      "_________________________________________________________________\n",
      "dropout_3 (Dropout)          (None, 10, 30)            0         \n",
      "_________________________________________________________________\n",
      "lstm_4 (LSTM)                (None, 10, 30)            7320      \n",
      "_________________________________________________________________\n",
      "dropout_4 (Dropout)          (None, 10, 30)            0         \n",
      "_________________________________________________________________\n",
      "lstm_5 (LSTM)                (None, 30)                7320      \n",
      "_________________________________________________________________\n",
      "dropout_5 (Dropout)          (None, 30)                0         \n",
      "_________________________________________________________________\n",
      "dense_1 (Dense)              (None, 1)                 31        \n",
      "=================================================================\n",
      "Total params: 18,511\n",
      "Trainable params: 18,511\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "metadata": {}
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "source": [
    "# Train the model\n",
    "# Use at least 10 epochs\n",
    "# Do not shuffle the data\n",
    "# Experiement with the batch size, but a smaller batch size is recommended\n",
    "\n",
    "model.fit(X_train, y_train, epochs=10, shuffle=False, batch_size=3, verbose=1)"
   ],
   "outputs": [
    {
     "output_type": "stream",
     "name": "stdout",
     "text": [
      "Epoch 1/10\n",
      "347/347 [==============================] - 3s 7ms/step - loss: 0.0025\n",
      "Epoch 2/10\n",
      "347/347 [==============================] - 3s 7ms/step - loss: 0.0021\n",
      "Epoch 3/10\n",
      "347/347 [==============================] - 3s 8ms/step - loss: 0.0021\n",
      "Epoch 4/10\n",
      "347/347 [==============================] - 3s 8ms/step - loss: 0.0020\n",
      "Epoch 5/10\n",
      "347/347 [==============================] - 3s 7ms/step - loss: 0.0018\n",
      "Epoch 6/10\n",
      "347/347 [==============================] - 3s 7ms/step - loss: 0.0020\n",
      "Epoch 7/10\n",
      "347/347 [==============================] - 2s 7ms/step - loss: 0.0016\n",
      "Epoch 8/10\n",
      "347/347 [==============================] - 3s 7ms/step - loss: 0.0018\n",
      "Epoch 9/10\n",
      "347/347 [==============================] - 3s 7ms/step - loss: 0.0015\n",
      "Epoch 10/10\n",
      "347/347 [==============================] - 3s 7ms/step - loss: 0.0016\n"
     ]
    },
    {
     "output_type": "execute_result",
     "data": {
      "text/plain": [
       "<tensorflow.python.keras.callbacks.History at 0x7f9ab4b69150>"
      ]
     },
     "metadata": {},
     "execution_count": 34
    }
   ],
   "metadata": {}
  },
  {
   "cell_type": "markdown",
   "source": [
    "---"
   ],
   "metadata": {}
  },
  {
   "cell_type": "markdown",
   "source": [
    "## Model Performance\n",
    "\n",
    "In this section, we will evaluate the model using the test data. \n",
    "\n",
    "Need to:\n",
    "1. Evaluate the model using the `X_test` and `y_test` data.\n",
    "2. Use the X_test data to make predictions\n",
    "3. Create a DataFrame of Real (y_test) vs predicted values. \n",
    "4. Plot the Real vs predicted values as a line chart\n",
    "\n",
    "### Hints\n",
    "Remember to apply the `inverse_transform` function to the predicted and y_test values to recover the actual closing prices."
   ],
   "metadata": {}
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "source": [
    "# Evaluate the model\n",
    "model.evaluate(X_test, y_test)"
   ],
   "outputs": [
    {
     "output_type": "stream",
     "name": "stdout",
     "text": [
      "14/14 [==============================] - 0s 3ms/step - loss: 0.0424\n"
     ]
    },
    {
     "output_type": "execute_result",
     "data": {
      "text/plain": [
       "0.04244405776262283"
      ]
     },
     "metadata": {},
     "execution_count": 35
    }
   ],
   "metadata": {}
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "source": [
    "# Make predictions\n",
    "predicted = model.predict(X_test)"
   ],
   "outputs": [],
   "metadata": {}
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "source": [
    "# Recover the original prices instead of the scaled version\n",
    "predicted_prices = scaler.inverse_transform(predicted)\n",
    "real_prices = scaler.inverse_transform(y_test.reshape(-1, 1))"
   ],
   "outputs": [],
   "metadata": {}
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "source": [
    "# Create a DataFrame of Real and Predicted values\n",
    "stock = pd.DataFrame({\"Real\": real_prices.ravel(),\"Predicted\": predicted_prices.ravel()}, index = stock_df.index[-len(real_prices): ]) \n",
    "stock.iloc[-20:]"
   ],
   "outputs": [
    {
     "output_type": "execute_result",
     "data": {
      "text/plain": [
       "                  Real   Predicted\n",
       "Date                              \n",
       "2021-06-21  274.040009  207.084244\n",
       "2021-06-22  277.100006  207.092178\n",
       "2021-06-23  276.920013  207.085373\n",
       "2021-06-24  274.660004  207.081207\n",
       "2021-06-25  275.489990  207.084045\n",
       "2021-06-28  278.380005  207.083542\n",
       "2021-06-29  276.609985  207.078964\n",
       "2021-06-30  276.209991  207.081207\n",
       "2021-07-01  277.920013  207.083984\n",
       "2021-07-02  280.170013  207.081772\n",
       "2021-07-06  279.119995  207.077744\n",
       "2021-07-07  277.320007  207.079407\n",
       "2021-07-08  279.239990  207.084167\n",
       "2021-07-09  274.100006  207.081879\n",
       "2021-07-12  280.489990  207.087830\n",
       "2021-07-13  280.109985  207.079773\n",
       "2021-07-14  279.209991  207.076477\n",
       "2021-07-15  278.309998  207.079849\n",
       "2021-07-16  279.720001  207.082901\n",
       "2021-07-19  277.899994  207.080978"
      ],
      "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>Real</th>\n",
       "      <th>Predicted</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Date</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2021-06-21</th>\n",
       "      <td>274.040009</td>\n",
       "      <td>207.084244</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2021-06-22</th>\n",
       "      <td>277.100006</td>\n",
       "      <td>207.092178</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2021-06-23</th>\n",
       "      <td>276.920013</td>\n",
       "      <td>207.085373</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2021-06-24</th>\n",
       "      <td>274.660004</td>\n",
       "      <td>207.081207</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2021-06-25</th>\n",
       "      <td>275.489990</td>\n",
       "      <td>207.084045</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2021-06-28</th>\n",
       "      <td>278.380005</td>\n",
       "      <td>207.083542</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2021-06-29</th>\n",
       "      <td>276.609985</td>\n",
       "      <td>207.078964</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2021-06-30</th>\n",
       "      <td>276.209991</td>\n",
       "      <td>207.081207</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2021-07-01</th>\n",
       "      <td>277.920013</td>\n",
       "      <td>207.083984</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2021-07-02</th>\n",
       "      <td>280.170013</td>\n",
       "      <td>207.081772</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2021-07-06</th>\n",
       "      <td>279.119995</td>\n",
       "      <td>207.077744</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2021-07-07</th>\n",
       "      <td>277.320007</td>\n",
       "      <td>207.079407</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2021-07-08</th>\n",
       "      <td>279.239990</td>\n",
       "      <td>207.084167</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2021-07-09</th>\n",
       "      <td>274.100006</td>\n",
       "      <td>207.081879</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2021-07-12</th>\n",
       "      <td>280.489990</td>\n",
       "      <td>207.087830</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2021-07-13</th>\n",
       "      <td>280.109985</td>\n",
       "      <td>207.079773</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2021-07-14</th>\n",
       "      <td>279.209991</td>\n",
       "      <td>207.076477</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2021-07-15</th>\n",
       "      <td>278.309998</td>\n",
       "      <td>207.079849</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2021-07-16</th>\n",
       "      <td>279.720001</td>\n",
       "      <td>207.082901</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2021-07-19</th>\n",
       "      <td>277.899994</td>\n",
       "      <td>207.080978</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ]
     },
     "metadata": {},
     "execution_count": 38
    }
   ],
   "metadata": {}
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "source": [
    "# Plot the real vs predicted values as a line chart\n",
    "stock.plot(rot=90)"
   ],
   "outputs": [
    {
     "output_type": "execute_result",
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x7f9ab4c4f690>"
      ]
     },
     "metadata": {},
     "execution_count": 39
    },
    {
     "output_type": "display_data",
     "data": {
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ],
      "image/png": "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"
     },
     "metadata": {
      "needs_background": "light"
     }
    }
   ],
   "metadata": {}
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "source": [
    "# In this case LSTM model is not reliable predictor of future stock prices aside from the period from 04/20-09/20 which was the time when coronavirus pandemic started. We need to tweak hyperparameters and length of history and compare the results, to see if we can improve the model without overfitting."
   ],
   "outputs": [],
   "metadata": {}
  }
 ],
 "metadata": {
  "file_extension": ".py",
  "kernelspec": {
   "name": "python3",
   "display_name": "Python 3.7.7 64-bit ('pyvizenv': conda)"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.7"
  },
  "mimetype": "text/x-python",
  "name": "python",
  "npconvert_exporter": "python",
  "pygments_lexer": "ipython3",
  "version": 3,
  "interpreter": {
   "hash": "6d180c226e1fb546a2a1623a132f2a8812479d52f4604fddc733d95b32da65e3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}