10 - Lesson - Recurrent Neural Network.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "4993ca55",
   "metadata": {},
   "source": [
    "# Recurrent Neural Network (RNN)\n",
    "### Goal of lesson\n",
    "- Understand Recurrent Neural Network (RNN)\n",
    "- Build a RNN on a timeseries\n",
    "- Hover over the theory of RNN (LSTM cells)\n",
    "- Use the MinMaxScaler from sklearn.\n",
    "- Create a RNN model with tensorflow\n",
    "- Applying the Dropout techniques.\n",
    "- Predict stock prices and make weather forecast using RNN.\n",
    "\n",
    "### Feed-Forward Neural Network\n",
    "- Neural network that has connection only in one direction\n",
    "\n",
    "### Recurrent Neural Network\n",
    "- Neural network that generates output that feeds back into its own inputs\n",
    "- Can do one-to-many and many-to-many relationship (not possible for feed-forward neural networks)\n",
    "- This can generate sentences (while feed-forward neural network can only generate \"words\" or fixed sets of outputs)\n",
    "- For time series\n",
    "\n",
    "![RNN](img/rnn.png)\n",
    "\n",
    "### Examples Recurrent Neural Network\n",
    "- Google translate\n",
    "- Voice recognition\n",
    "- Video copy right violation"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "11047afa",
   "metadata": {},
   "source": [
    "### Theory\n",
    "- It is complex - and luckily - not needed to understand in depth\n",
    "- Quote of the author of Keras (Francios Chollet):\n",
    "    - \"*You don't need to understand everything about the specific architecutre of an LSTM cell [...] just that LSTM cell is meant to allow past information to be reinjected at a later time.*\""
   ]
  },
  {
   "cell_type": "markdown",
   "id": "61c11de8",
   "metadata": {},
   "source": [
    "> #### Programming Notes:\n",
    "> - Libraries used\n",
    ">     - [**pandas**](https://pandas.pydata.org) - a data analysis and manipulation tool\n",
    ">     - [**numpy**](http://numpy.org) - scientific computing with Python ([Lecture on NumPy](https://youtu.be/BpzpU8_j0-c))\n",
    ">     - [**matplotlib**](http://matplotlib.org) - visualization with Python ([Lecture on **visualization**](https://youtu.be/htIh8YHh4xs))\n",
    ">     - [**tensorflow**](https://www.tensorflow.org) - end-to-end open source machine learning platform\n",
    "> - Functionality and concepts used\n",
    ">     - [**CSV**](https://en.wikipedia.org/wiki/Comma-separated_values) file ([Lecture on CSV](https://youtu.be/LEyojSOg4EI))\n",
    ">     - [**read_csv()**](https://pandas.pydata.org/pandas-docs/stable/reference/api/pandas.read_csv.html) read a comma-separated values (csv) file into **pandas** DataFrame.\n",
    ">     - [**MinMaxScaler**](https://scikit-learn.org/stable/modules/generated/sklearn.preprocessing.MinMaxScaler.html) Transform features by scaling each feature to a given range\n",
    ">     - [**Sequential**](https://www.tensorflow.org/api_docs/python/tf/keras/Sequential) provides training and inference features for a model\n",
    ">     - [**Dense**](https://www.tensorflow.org/api_docs/python/tf/keras/layers/Dense) regular densely-connected Neural Network (NN) layer.\n",
    ">     - [**Dropout**](https://www.tensorflow.org/api_docs/python/tf/keras/layers/Dropout) applies Dropout to the input\n",
    ">     - [**LSTM**](https://www.tensorflow.org/api_docs/python/tf/keras/layers/LSTM) Long Short-Term Memory layer - Hochreiter 1997"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "c50eceef",
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "from sklearn.preprocessing import MinMaxScaler\n",
    "from tensorflow.keras.models import Sequential\n",
    "from tensorflow.keras.layers import Dense\n",
    "from tensorflow.keras.layers import LSTM\n",
    "from tensorflow.keras.layers import Dropout\n",
    "import matplotlib.pyplot as plt\n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "1915b5d5",
   "metadata": {},
   "outputs": [],
   "source": [
    "data = pd.read_csv('files/aapl.csv', parse_dates=True, index_col=0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "30a00a53",
   "metadata": {},
   "outputs": [
    {
     "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>High</th>\n",
       "      <th>Low</th>\n",
       "      <th>Open</th>\n",
       "      <th>Close</th>\n",
       "      <th>Volume</th>\n",
       "      <th>Adj Close</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Date</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2000-01-03</th>\n",
       "      <td>1.004464</td>\n",
       "      <td>0.907924</td>\n",
       "      <td>0.936384</td>\n",
       "      <td>0.999442</td>\n",
       "      <td>535796800.0</td>\n",
       "      <td>0.858137</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2000-01-04</th>\n",
       "      <td>0.987723</td>\n",
       "      <td>0.903460</td>\n",
       "      <td>0.966518</td>\n",
       "      <td>0.915179</td>\n",
       "      <td>512377600.0</td>\n",
       "      <td>0.785788</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2000-01-05</th>\n",
       "      <td>0.987165</td>\n",
       "      <td>0.919643</td>\n",
       "      <td>0.926339</td>\n",
       "      <td>0.928571</td>\n",
       "      <td>778321600.0</td>\n",
       "      <td>0.797286</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2000-01-06</th>\n",
       "      <td>0.955357</td>\n",
       "      <td>0.848214</td>\n",
       "      <td>0.947545</td>\n",
       "      <td>0.848214</td>\n",
       "      <td>767972800.0</td>\n",
       "      <td>0.728291</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2000-01-07</th>\n",
       "      <td>0.901786</td>\n",
       "      <td>0.852679</td>\n",
       "      <td>0.861607</td>\n",
       "      <td>0.888393</td>\n",
       "      <td>460734400.0</td>\n",
       "      <td>0.762789</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                High       Low      Open     Close       Volume  Adj Close\n",
       "Date                                                                      \n",
       "2000-01-03  1.004464  0.907924  0.936384  0.999442  535796800.0   0.858137\n",
       "2000-01-04  0.987723  0.903460  0.966518  0.915179  512377600.0   0.785788\n",
       "2000-01-05  0.987165  0.919643  0.926339  0.928571  778321600.0   0.797286\n",
       "2000-01-06  0.955357  0.848214  0.947545  0.848214  767972800.0   0.728291\n",
       "2000-01-07  0.901786  0.852679  0.861607  0.888393  460734400.0   0.762789"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "49353d97",
   "metadata": {},
   "outputs": [],
   "source": [
    "data_train = data.loc['2000':'2019', 'Adj Close'].to_numpy()\n",
    "data_test = data.loc['2020', 'Adj Close'].to_numpy()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "96841622",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([ 0.85813743,  0.78578764,  0.79728615, ..., 71.49375916,\n",
       "       71.91808319, 72.44355774])"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data_train"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "9f7b4674",
   "metadata": {},
   "outputs": [],
   "source": [
    "scaler = MinMaxScaler()\n",
    "data_train = scaler.fit_transform(data_train.reshape(-1, 1))\n",
    "data_test = scaler.transform(data_test.reshape(-1, 1))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "734aa56e",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[0.00909405],\n",
       "       [0.00809256],\n",
       "       [0.00825172],\n",
       "       ...,\n",
       "       [0.98685262],\n",
       "       [0.99272623],\n",
       "       [1.        ]])"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data_train"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "a5618852",
   "metadata": {},
   "outputs": [],
   "source": [
    "def data_preparation(data):\n",
    "    x = []\n",
    "    y = []\n",
    "    \n",
    "    for i in range(40, len(data)):\n",
    "        x.append(data[i-40:i, 0])\n",
    "        y.append(data[i])\n",
    "        \n",
    "    x = np.array(x)\n",
    "    y = np.array(y)\n",
    "    \n",
    "    x = x.reshape(x.shape[0], x.shape[1], 1)\n",
    "    \n",
    "    return x, y"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "35b64d05",
   "metadata": {},
   "outputs": [],
   "source": [
    "x_train, y_train = data_preparation(data_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "45ab98c5",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "((4991, 40, 1), (4991, 1))"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "x_train.shape, y_train.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "ab6c898d",
   "metadata": {},
   "outputs": [],
   "source": [
    "x_test, y_test = data_preparation(data_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "dc01e6cf",
   "metadata": {},
   "outputs": [],
   "source": [
    "model = Sequential()\n",
    "model.add(LSTM(units=45, return_sequences=True, input_shape=(x_train.shape[1], 1)))\n",
    "model.add(LSTM(units=45, return_sequences=True))\n",
    "model.add(LSTM(units=45))\n",
    "model.add(Dense(units=1))\n",
    "\n",
    "model.compile(optimizer='adam', loss='mean_squared_error', metrics=['accuracy'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "id": "ad492a17",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/5\n",
      "156/156 [==============================] - 7s 28ms/step - loss: 0.0038 - accuracy: 4.0072e-04\n",
      "Epoch 2/5\n",
      "156/156 [==============================] - 4s 27ms/step - loss: 2.4621e-04 - accuracy: 4.0072e-04\n",
      "Epoch 3/5\n",
      "156/156 [==============================] - 4s 28ms/step - loss: 1.9201e-04 - accuracy: 4.0072e-04\n",
      "Epoch 4/5\n",
      "156/156 [==============================] - 4s 28ms/step - loss: 1.7177e-04 - accuracy: 4.0072e-04\n",
      "Epoch 5/5\n",
      "156/156 [==============================] - 4s 28ms/step - loss: 1.4874e-04 - accuracy: 4.0072e-04\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<tensorflow.python.keras.callbacks.History at 0x7fa44c442fa0>"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.fit(x_train, y_train, epochs=5, batch_size=32)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "c258ab6e",
   "metadata": {},
   "outputs": [],
   "source": [
    "y_pred = model.predict(x_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "id": "9d3fa5cc",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(213, 1)"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_pred.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "id": "fd0bcae1",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[<matplotlib.lines.Line2D at 0x7fa43929d610>]"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "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": [
    "fig, ax = plt.subplots()\n",
    "\n",
    "ax.plot(y_pred)\n",
    "ax.plot(y_test, c='r')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "id": "0eed5e1b",
   "metadata": {},
   "outputs": [],
   "source": [
    "y_unscaled = scaler.inverse_transform(y_pred)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "id": "8d5f8c02",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[<matplotlib.lines.Line2D at 0x7fa44f694700>]"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "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": [
    "fig, ax = plt.subplots()\n",
    "\n",
    "y_real = data.loc['2020', 'Adj Close'].to_numpy()\n",
    "\n",
    "ax.plot(y_real[40:])\n",
    "ax.plot(y_unscaled)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "0274c06f",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "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": "3.8.8"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}