Keras - Time series prediction with LSTM.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/abulbasar/anaconda3/lib/python3.6/site-packages/h5py/__init__.py:36: FutureWarning: Conversion of the second argument of issubdtype from `float` to `np.floating` is deprecated. In future, it will be treated as `np.float64 == np.dtype(float).type`.\n",
      "  from ._conv import register_converters as _register_converters\n",
      "Using TensorFlow backend.\n"
     ]
    }
   ],
   "source": [
    "import keras\n",
    "from keras import Sequential\n",
    "from keras.layers import Dense, Dropout, LSTM\n",
    "from keras.callbacks import TensorBoard\n",
    "import tensorflow as tf\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "\n",
    "from time import time\n",
    "\n",
    "import matplotlib.pyplot as plt"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "('2.1.5', '1.6.0', '1.14.2')"
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "keras.__version__, tf.__version__, np.__version__"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([112, 118, 132, 129, 121, 135, 148, 148, 136, 119, 104, 118, 115,\n",
       "       126, 141, 135, 125, 149, 170, 170, 158, 133, 114, 140, 145, 150,\n",
       "       178, 163, 172, 178, 199, 199, 184, 162, 146, 166, 171, 180, 193,\n",
       "       181, 183, 218, 230, 242, 209, 191, 172, 194, 196, 196, 236, 235,\n",
       "       229, 243, 264, 272, 237, 211, 180, 201, 204, 188, 235, 227, 234,\n",
       "       264, 302, 293, 259, 229, 203, 229, 242, 233, 267, 269, 270, 315,\n",
       "       364, 347, 312, 274, 237, 278, 284, 277, 317, 313, 318, 374, 413,\n",
       "       405, 355, 306, 271, 306, 315, 301, 356, 348, 355, 422, 465, 467,\n",
       "       404, 347, 305, 336, 340, 318, 362, 348, 363, 435, 491, 505, 404,\n",
       "       359, 310, 337, 360, 342, 406, 396, 420, 472, 548, 559, 463, 407,\n",
       "       362, 405, 417, 391, 419, 461, 472, 535, 622, 606, 508, 461, 390])"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data = pd.read_csv(\"/data/airline-passengers.csv\", skiprows=1\n",
    "                , skipfooter=4, names=[\"date\", \"passenger_count\"], engine=\"python\").passenger_count.values\n",
    "data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[<matplotlib.lines.Line2D at 0x18268136d8>]"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x181b1ccd68>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(data)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "count    143.000000\n",
       "mean     279.237762\n",
       "std      119.708270\n",
       "min      104.000000\n",
       "25%      180.000000\n",
       "50%      264.000000\n",
       "75%      359.500000\n",
       "max      622.000000\n",
       "dtype: float64"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pd.Series(data).describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([0.01544402, 0.02702703, 0.05405405, 0.04826255, 0.03281853,\n",
       "       0.05984556, 0.08494208, 0.08494208, 0.06177606, 0.02895753,\n",
       "       0.        , 0.02702703, 0.02123552, 0.04247104, 0.07142857,\n",
       "       0.05984556, 0.04054054, 0.08687259, 0.12741313, 0.12741313,\n",
       "       0.1042471 , 0.05598456, 0.01930502, 0.06949807, 0.07915058,\n",
       "       0.08880309, 0.14285714, 0.11389961, 0.13127413, 0.14285714,\n",
       "       0.18339768, 0.18339768, 0.15444015, 0.11196911, 0.08108108,\n",
       "       0.11969112, 0.12934363, 0.14671815, 0.17181467, 0.14864865,\n",
       "       0.15250965, 0.22007722, 0.24324324, 0.26640927, 0.2027027 ,\n",
       "       0.16795367, 0.13127413, 0.17374517, 0.17760618, 0.17760618,\n",
       "       0.25482625, 0.25289575, 0.24131274, 0.26833977, 0.30888031,\n",
       "       0.32432432, 0.25675676, 0.20656371, 0.14671815, 0.18725869,\n",
       "       0.19305019, 0.16216216, 0.25289575, 0.23745174, 0.25096525,\n",
       "       0.30888031, 0.38223938, 0.36486486, 0.2992278 , 0.24131274,\n",
       "       0.19111969, 0.24131274, 0.26640927, 0.24903475, 0.31467181,\n",
       "       0.31853282, 0.32046332, 0.40733591, 0.5019305 , 0.46911197,\n",
       "       0.4015444 , 0.32818533, 0.25675676, 0.33590734, 0.34749035,\n",
       "       0.33397683, 0.41119691, 0.4034749 , 0.41312741, 0.52123552,\n",
       "       0.5965251 , 0.58108108, 0.48455598, 0.38996139, 0.32239382,\n",
       "       0.38996139, 0.40733591, 0.38030888, 0.48648649, 0.47104247,\n",
       "       0.48455598, 0.61389961, 0.6969112 , 0.7007722 , 0.57915058,\n",
       "       0.46911197, 0.38803089, 0.44787645, 0.45559846, 0.41312741,\n",
       "       0.4980695 , 0.47104247, 0.5       , 0.63899614, 0.74710425,\n",
       "       0.77413127, 0.57915058, 0.49227799, 0.3976834 , 0.44980695,\n",
       "       0.49420849, 0.45945946, 0.58301158, 0.56370656, 0.61003861,\n",
       "       0.71042471, 0.85714286, 0.87837838, 0.69305019, 0.58494208,\n",
       "       0.4980695 , 0.58108108, 0.6042471 , 0.55405405, 0.60810811,\n",
       "       0.68918919, 0.71042471, 0.83204633, 1.        , 0.96911197,\n",
       "       0.77992278, 0.68918919, 0.55212355])"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data_scaled = (data - np.min(data)) / (np.max(data) - np.min(data))\n",
    "data_scaled"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[<matplotlib.lines.Line2D at 0x182690fc88>]"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x181b2a77f0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(data_scaled)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "def rmse(y_true, y_pred):\n",
    "    return np.sqrt(np.mean((y_true - y_pred) ** 2))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "training_size = int(len(data) * 0.7)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "training = data_scaled[:training_size]\n",
    "testing = data_scaled[training_size: ]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "def create_sequences(data, sequence_len, step = 1):\n",
    "    data_len = len(data)\n",
    "    X = []\n",
    "    y = []\n",
    "    for start in range(0, data_len - sequence_len, step):\n",
    "        end = start + sequence_len\n",
    "        sequence = data[start: end]\n",
    "        X.append(sequence)\n",
    "        y.append(data[end])\n",
    "    return np.array(X), np.array(y)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "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>x0</th>\n",
       "      <th>x1</th>\n",
       "      <th>x2</th>\n",
       "      <th>x3</th>\n",
       "      <th>x4</th>\n",
       "      <th>y</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.015444</td>\n",
       "      <td>0.027027</td>\n",
       "      <td>0.054054</td>\n",
       "      <td>0.048263</td>\n",
       "      <td>0.032819</td>\n",
       "      <td>0.059846</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.027027</td>\n",
       "      <td>0.054054</td>\n",
       "      <td>0.048263</td>\n",
       "      <td>0.032819</td>\n",
       "      <td>0.059846</td>\n",
       "      <td>0.084942</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.054054</td>\n",
       "      <td>0.048263</td>\n",
       "      <td>0.032819</td>\n",
       "      <td>0.059846</td>\n",
       "      <td>0.084942</td>\n",
       "      <td>0.084942</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.048263</td>\n",
       "      <td>0.032819</td>\n",
       "      <td>0.059846</td>\n",
       "      <td>0.084942</td>\n",
       "      <td>0.084942</td>\n",
       "      <td>0.061776</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.032819</td>\n",
       "      <td>0.059846</td>\n",
       "      <td>0.084942</td>\n",
       "      <td>0.084942</td>\n",
       "      <td>0.061776</td>\n",
       "      <td>0.028958</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "         x0        x1        x2        x3        x4         y\n",
       "0  0.015444  0.027027  0.054054  0.048263  0.032819  0.059846\n",
       "1  0.027027  0.054054  0.048263  0.032819  0.059846  0.084942\n",
       "2  0.054054  0.048263  0.032819  0.059846  0.084942  0.084942\n",
       "3  0.048263  0.032819  0.059846  0.084942  0.084942  0.061776\n",
       "4  0.032819  0.059846  0.084942  0.084942  0.061776  0.028958"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X, y = create_sequences(data_scaled, sequence_len= 5)\n",
    "df = pd.DataFrame(X, columns=[\"x%d\" % i for i in range(X.shape[1])])\n",
    "df[\"y\"] = y\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "scrolled": false
   },
   "outputs": [],
   "source": [
    "def show_values(y_train, y_test, y_train_pred, y_test_pred):\n",
    "    y_train_pred = y_train_pred.flatten()\n",
    "    y_test_pred = y_test_pred.flatten()\n",
    "    plt.plot(np.append(y_train, y_test), color = \"lightblue\", label = \"Actual\")\n",
    "    plt.plot(range(0, len(y_train_pred)), y_train_pred, color=\"orange\", label = \"Prediction - train\")\n",
    "    plt.plot(range(len(y_train_pred), len(y_train_pred) + len(y_test_pred)), y_test_pred, color=\"red\", label = \"Prediction - test\")\n",
    "    plt.legend()\n",
    "    train_rmse = rmse(y_train, y_train_pred)\n",
    "    test_rmse = rmse(y_test, y_test_pred)\n",
    "    plt.title(\"RMSE (train: %f, test: %f)\" % (train_rmse, test_rmse))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "lstm_1 (LSTM)                (None, 4)                 96        \n",
      "_________________________________________________________________\n",
      "dense_1 (Dense)              (None, 1)                 5         \n",
      "=================================================================\n",
      "Total params: 101\n",
      "Trainable params: 101\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n",
      "Epoch 1/100\n",
      "99/99 [==============================] - 1s 6ms/step - loss: 0.0460\n",
      "Epoch 2/100\n",
      "99/99 [==============================] - 0s 4ms/step - loss: 0.0209\n",
      "Epoch 3/100\n",
      "99/99 [==============================] - 0s 3ms/step - loss: 0.0162\n",
      "Epoch 4/100\n",
      "99/99 [==============================] - 0s 3ms/step - loss: 0.0149\n",
      "Epoch 5/100\n",
      "99/99 [==============================] - 0s 4ms/step - loss: 0.0138\n",
      "Epoch 6/100\n",
      "99/99 [==============================] - 0s 4ms/step - loss: 0.0130\n",
      "Epoch 7/100\n",
      "99/99 [==============================] - 0s 5ms/step - loss: 0.0116\n",
      "Epoch 8/100\n",
      "99/99 [==============================] - 0s 4ms/step - loss: 0.0105\n",
      "Epoch 9/100\n",
      "99/99 [==============================] - 0s 4ms/step - loss: 0.0095\n",
      "Epoch 10/100\n",
      "99/99 [==============================] - 0s 4ms/step - loss: 0.0084\n",
      "Epoch 11/100\n",
      "99/99 [==============================] - 0s 4ms/step - loss: 0.0076\n",
      "Epoch 12/100\n",
      "99/99 [==============================] - 0s 4ms/step - loss: 0.0067\n",
      "Epoch 13/100\n",
      "99/99 [==============================] - 0s 3ms/step - loss: 0.0059\n",
      "Epoch 14/100\n",
      "99/99 [==============================] - 0s 3ms/step - loss: 0.0051\n",
      "Epoch 15/100\n",
      "99/99 [==============================] - 0s 4ms/step - loss: 0.0046\n",
      "Epoch 16/100\n",
      "99/99 [==============================] - 0s 3ms/step - loss: 0.0039\n",
      "Epoch 17/100\n",
      "99/99 [==============================] - 0s 4ms/step - loss: 0.0035\n",
      "Epoch 18/100\n",
      "99/99 [==============================] - 0s 3ms/step - loss: 0.0031\n",
      "Epoch 19/100\n",
      "99/99 [==============================] - 0s 4ms/step - loss: 0.0028\n",
      "Epoch 20/100\n",
      "99/99 [==============================] - 0s 4ms/step - loss: 0.0026\n",
      "Epoch 21/100\n",
      "99/99 [==============================] - 0s 4ms/step - loss: 0.0024\n",
      "Epoch 22/100\n",
      "99/99 [==============================] - 0s 3ms/step - loss: 0.0023\n",
      "Epoch 23/100\n",
      "99/99 [==============================] - 0s 4ms/step - loss: 0.0022\n",
      "Epoch 24/100\n",
      "99/99 [==============================] - 0s 3ms/step - loss: 0.0022\n",
      "Epoch 25/100\n",
      "99/99 [==============================] - 0s 3ms/step - loss: 0.0022\n",
      "Epoch 26/100\n",
      "99/99 [==============================] - 0s 4ms/step - loss: 0.0022\n",
      "Epoch 27/100\n",
      "99/99 [==============================] - 0s 4ms/step - loss: 0.0021\n",
      "Epoch 28/100\n",
      "99/99 [==============================] - 0s 4ms/step - loss: 0.0022\n",
      "Epoch 29/100\n",
      "99/99 [==============================] - 0s 4ms/step - loss: 0.0022\n",
      "Epoch 30/100\n",
      "99/99 [==============================] - 0s 4ms/step - loss: 0.0022\n",
      "Epoch 31/100\n",
      "99/99 [==============================] - 0s 4ms/step - loss: 0.0022\n",
      "Epoch 32/100\n",
      "99/99 [==============================] - 0s 4ms/step - loss: 0.0021\n",
      "Epoch 33/100\n",
      "99/99 [==============================] - 0s 4ms/step - loss: 0.0021\n",
      "Epoch 34/100\n",
      "99/99 [==============================] - 0s 4ms/step - loss: 0.0021\n",
      "Epoch 35/100\n",
      "99/99 [==============================] - 0s 4ms/step - loss: 0.0022\n",
      "Epoch 36/100\n",
      "99/99 [==============================] - 0s 3ms/step - loss: 0.0022\n",
      "Epoch 37/100\n",
      "99/99 [==============================] - 0s 4ms/step - loss: 0.0022\n",
      "Epoch 38/100\n",
      "99/99 [==============================] - 0s 4ms/step - loss: 0.0021\n",
      "Epoch 39/100\n",
      "99/99 [==============================] - 0s 4ms/step - loss: 0.0022\n",
      "Epoch 40/100\n",
      "99/99 [==============================] - 0s 4ms/step - loss: 0.0021\n",
      "Epoch 41/100\n",
      "99/99 [==============================] - 0s 3ms/step - loss: 0.0021\n",
      "Epoch 42/100\n",
      "99/99 [==============================] - 0s 4ms/step - loss: 0.0021\n",
      "Epoch 43/100\n",
      "99/99 [==============================] - 0s 4ms/step - loss: 0.0022\n",
      "Epoch 44/100\n",
      "99/99 [==============================] - 0s 4ms/step - loss: 0.0021\n",
      "Epoch 45/100\n",
      "99/99 [==============================] - 0s 3ms/step - loss: 0.0021\n",
      "Epoch 46/100\n",
      "99/99 [==============================] - 0s 4ms/step - loss: 0.0021\n",
      "Epoch 47/100\n",
      "99/99 [==============================] - 0s 3ms/step - loss: 0.0020\n",
      "Epoch 48/100\n",
      "99/99 [==============================] - 0s 4ms/step - loss: 0.0022\n",
      "Epoch 49/100\n",
      "99/99 [==============================] - 0s 3ms/step - loss: 0.0021\n",
      "Epoch 50/100\n",
      "99/99 [==============================] - 0s 3ms/step - loss: 0.0021\n",
      "Epoch 51/100\n",
      "99/99 [==============================] - 0s 4ms/step - loss: 0.0022\n",
      "Epoch 52/100\n",
      "99/99 [==============================] - 0s 4ms/step - loss: 0.0021\n",
      "Epoch 53/100\n",
      "99/99 [==============================] - 0s 4ms/step - loss: 0.0021\n",
      "Epoch 54/100\n",
      "99/99 [==============================] - 0s 4ms/step - loss: 0.0021\n",
      "Epoch 55/100\n",
      "99/99 [==============================] - 0s 4ms/step - loss: 0.0022\n",
      "Epoch 56/100\n",
      "99/99 [==============================] - 0s 4ms/step - loss: 0.0021\n",
      "Epoch 57/100\n",
      "99/99 [==============================] - 0s 3ms/step - loss: 0.0022\n",
      "Epoch 58/100\n",
      "99/99 [==============================] - 0s 4ms/step - loss: 0.0020\n",
      "Epoch 59/100\n",
      "99/99 [==============================] - 0s 4ms/step - loss: 0.0021\n",
      "Epoch 60/100\n",
      "99/99 [==============================] - 0s 4ms/step - loss: 0.0021\n",
      "Epoch 61/100\n",
      "99/99 [==============================] - 0s 3ms/step - loss: 0.0021\n",
      "Epoch 62/100\n",
      "99/99 [==============================] - 0s 4ms/step - loss: 0.0021\n",
      "Epoch 63/100\n",
      "99/99 [==============================] - 0s 3ms/step - loss: 0.0022\n",
      "Epoch 64/100\n",
      "99/99 [==============================] - 0s 3ms/step - loss: 0.0021\n",
      "Epoch 65/100\n",
      "99/99 [==============================] - 0s 4ms/step - loss: 0.0021\n",
      "Epoch 66/100\n",
      "99/99 [==============================] - 0s 3ms/step - loss: 0.0021\n",
      "Epoch 67/100\n",
      "99/99 [==============================] - 0s 3ms/step - loss: 0.0020\n",
      "Epoch 68/100\n",
      "99/99 [==============================] - 0s 4ms/step - loss: 0.0021\n",
      "Epoch 69/100\n",
      "99/99 [==============================] - 0s 3ms/step - loss: 0.0021\n",
      "Epoch 70/100\n",
      "99/99 [==============================] - 0s 4ms/step - loss: 0.0023\n",
      "Epoch 71/100\n",
      "99/99 [==============================] - 0s 3ms/step - loss: 0.0021\n",
      "Epoch 72/100\n",
      "99/99 [==============================] - 0s 4ms/step - loss: 0.0021\n",
      "Epoch 73/100\n",
      "99/99 [==============================] - 0s 3ms/step - loss: 0.0021\n",
      "Epoch 74/100\n",
      "99/99 [==============================] - 0s 3ms/step - loss: 0.0021\n",
      "Epoch 75/100\n",
      "99/99 [==============================] - 0s 4ms/step - loss: 0.0021\n",
      "Epoch 76/100\n",
      "99/99 [==============================] - 0s 4ms/step - loss: 0.0022\n",
      "Epoch 77/100\n",
      "99/99 [==============================] - 0s 4ms/step - loss: 0.0021\n",
      "Epoch 78/100\n",
      "99/99 [==============================] - 0s 4ms/step - loss: 0.0021\n",
      "Epoch 79/100\n",
      "99/99 [==============================] - 0s 3ms/step - loss: 0.0021\n",
      "Epoch 80/100\n",
      "99/99 [==============================] - 0s 3ms/step - loss: 0.0020\n",
      "Epoch 81/100\n",
      "99/99 [==============================] - 0s 3ms/step - loss: 0.0021\n",
      "Epoch 82/100\n",
      "99/99 [==============================] - 0s 3ms/step - loss: 0.0021\n",
      "Epoch 83/100\n",
      "99/99 [==============================] - 0s 4ms/step - loss: 0.0021\n",
      "Epoch 84/100\n",
      "99/99 [==============================] - 0s 4ms/step - loss: 0.0020\n",
      "Epoch 85/100\n",
      "99/99 [==============================] - 0s 4ms/step - loss: 0.0021\n",
      "Epoch 86/100\n",
      "99/99 [==============================] - 0s 4ms/step - loss: 0.0021\n",
      "Epoch 87/100\n",
      "99/99 [==============================] - 0s 4ms/step - loss: 0.0021\n",
      "Epoch 88/100\n",
      "99/99 [==============================] - 0s 4ms/step - loss: 0.0021\n",
      "Epoch 89/100\n",
      "99/99 [==============================] - 0s 4ms/step - loss: 0.0021\n",
      "Epoch 90/100\n",
      "99/99 [==============================] - 0s 4ms/step - loss: 0.0021\n",
      "Epoch 91/100\n",
      "99/99 [==============================] - 0s 3ms/step - loss: 0.0021\n",
      "Epoch 92/100\n",
      "99/99 [==============================] - 0s 3ms/step - loss: 0.0021\n",
      "Epoch 93/100\n",
      "99/99 [==============================] - 0s 4ms/step - loss: 0.0021\n",
      "Epoch 94/100\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "99/99 [==============================] - 1s 6ms/step - loss: 0.0021\n",
      "Epoch 95/100\n",
      "99/99 [==============================] - 0s 3ms/step - loss: 0.0020\n",
      "Epoch 96/100\n",
      "99/99 [==============================] - 0s 4ms/step - loss: 0.0022\n",
      "Epoch 97/100\n",
      "99/99 [==============================] - 0s 3ms/step - loss: 0.0021\n",
      "Epoch 98/100\n",
      "99/99 [==============================] - 0s 3ms/step - loss: 0.0021\n",
      "Epoch 99/100\n",
      "99/99 [==============================] - 0s 3ms/step - loss: 0.0021\n",
      "Epoch 100/100\n",
      "99/99 [==============================] - 0s 3ms/step - loss: 0.0021\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1c2a01f860>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "sequence_len = 1\n",
    "X_train, y_train = create_sequences(training, sequence_len)\n",
    "X_test, y_test = create_sequences(testing, sequence_len)\n",
    "\n",
    "model = Sequential()\n",
    "model.add(LSTM(4, activation=\"tanh\", input_shape=(sequence_len, 1)))\n",
    "model.add(Dense(1, activation=None))\n",
    "model.compile(loss='mean_squared_error', optimizer='adam')\n",
    "model.summary()\n",
    "\n",
    "tensor_board = TensorBoard(log_dir=\"/tmp/tf/logs/%d\" % time())\n",
    "model.fit(X_train.reshape(-1, sequence_len, 1), y_train\n",
    "          , epochs=100\n",
    "          , batch_size=1\n",
    "          , verbose=1\n",
    "          , callbacks=[tensor_board])\n",
    "\n",
    "y_train_pred = model.predict(X_train.reshape(-1, sequence_len, 1))\n",
    "y_test_pred = model.predict(X_test.reshape(-1, sequence_len, 1))\n",
    "show_values(y_train, y_test, y_train_pred, y_test_pred)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "lstm_2 (LSTM)                (None, 4)                 96        \n",
      "_________________________________________________________________\n",
      "dense_2 (Dense)              (None, 1)                 5         \n",
      "=================================================================\n",
      "Total params: 101\n",
      "Trainable params: 101\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n",
      "Epoch 1/100\n",
      "97/97 [==============================] - 1s 7ms/step - loss: 0.0265\n",
      "Epoch 2/100\n",
      "97/97 [==============================] - 1s 6ms/step - loss: 0.0111\n",
      "Epoch 3/100\n",
      "97/97 [==============================] - 1s 6ms/step - loss: 0.0094\n",
      "Epoch 4/100\n",
      "97/97 [==============================] - 0s 5ms/step - loss: 0.0079\n",
      "Epoch 5/100\n",
      "97/97 [==============================] - 0s 4ms/step - loss: 0.0069\n",
      "Epoch 6/100\n",
      "97/97 [==============================] - 0s 5ms/step - loss: 0.0060\n",
      "Epoch 7/100\n",
      "97/97 [==============================] - 1s 6ms/step - loss: 0.0051\n",
      "Epoch 8/100\n",
      "97/97 [==============================] - 1s 6ms/step - loss: 0.0046\n",
      "Epoch 9/100\n",
      "97/97 [==============================] - 1s 7ms/step - loss: 0.0042\n",
      "Epoch 10/100\n",
      "97/97 [==============================] - 1s 6ms/step - loss: 0.0041\n",
      "Epoch 11/100\n",
      "97/97 [==============================] - 1s 7ms/step - loss: 0.0039\n",
      "Epoch 12/100\n",
      "97/97 [==============================] - 0s 4ms/step - loss: 0.0039\n",
      "Epoch 13/100\n",
      "97/97 [==============================] - 0s 5ms/step - loss: 0.0037\n",
      "Epoch 14/100\n",
      "97/97 [==============================] - 1s 8ms/step - loss: 0.0038\n",
      "Epoch 15/100\n",
      "97/97 [==============================] - 1s 7ms/step - loss: 0.0038\n",
      "Epoch 16/100\n",
      "97/97 [==============================] - 0s 5ms/step - loss: 0.0038\n",
      "Epoch 17/100\n",
      "97/97 [==============================] - 0s 4ms/step - loss: 0.0038\n",
      "Epoch 18/100\n",
      "97/97 [==============================] - 0s 4ms/step - loss: 0.0037\n",
      "Epoch 19/100\n",
      "97/97 [==============================] - 0s 4ms/step - loss: 0.0036\n",
      "Epoch 20/100\n",
      "97/97 [==============================] - 1s 6ms/step - loss: 0.0036\n",
      "Epoch 21/100\n",
      "97/97 [==============================] - 1s 6ms/step - loss: 0.0036\n",
      "Epoch 22/100\n",
      "97/97 [==============================] - 1s 7ms/step - loss: 0.0036\n",
      "Epoch 23/100\n",
      "97/97 [==============================] - 1s 5ms/step - loss: 0.0036\n",
      "Epoch 24/100\n",
      "97/97 [==============================] - 1s 5ms/step - loss: 0.0037\n",
      "Epoch 25/100\n",
      "97/97 [==============================] - 0s 4ms/step - loss: 0.0037\n",
      "Epoch 26/100\n",
      "97/97 [==============================] - 0s 4ms/step - loss: 0.0036\n",
      "Epoch 27/100\n",
      "97/97 [==============================] - 1s 7ms/step - loss: 0.0035\n",
      "Epoch 28/100\n",
      "97/97 [==============================] - 1s 6ms/step - loss: 0.0036\n",
      "Epoch 29/100\n",
      "97/97 [==============================] - 1s 7ms/step - loss: 0.0036\n",
      "Epoch 30/100\n",
      "97/97 [==============================] - 1s 6ms/step - loss: 0.0036\n",
      "Epoch 31/100\n",
      "97/97 [==============================] - 1s 5ms/step - loss: 0.0035\n",
      "Epoch 32/100\n",
      "97/97 [==============================] - 1s 6ms/step - loss: 0.0036\n",
      "Epoch 33/100\n",
      "97/97 [==============================] - 0s 5ms/step - loss: 0.0035\n",
      "Epoch 34/100\n",
      "97/97 [==============================] - 1s 7ms/step - loss: 0.0036\n",
      "Epoch 35/100\n",
      "97/97 [==============================] - 0s 5ms/step - loss: 0.0036\n",
      "Epoch 36/100\n",
      "97/97 [==============================] - 1s 6ms/step - loss: 0.0034\n",
      "Epoch 37/100\n",
      "97/97 [==============================] - 1s 7ms/step - loss: 0.0035\n",
      "Epoch 38/100\n",
      "97/97 [==============================] - 1s 5ms/step - loss: 0.0034\n",
      "Epoch 39/100\n",
      "97/97 [==============================] - 1s 8ms/step - loss: 0.0035\n",
      "Epoch 40/100\n",
      "97/97 [==============================] - 1s 5ms/step - loss: 0.0035\n",
      "Epoch 41/100\n",
      "97/97 [==============================] - 0s 5ms/step - loss: 0.0035\n",
      "Epoch 42/100\n",
      "97/97 [==============================] - 1s 8ms/step - loss: 0.0034\n",
      "Epoch 43/100\n",
      "97/97 [==============================] - 1s 12ms/step - loss: 0.0035\n",
      "Epoch 44/100\n",
      "97/97 [==============================] - 1s 8ms/step - loss: 0.0034\n",
      "Epoch 45/100\n",
      "97/97 [==============================] - 1s 7ms/step - loss: 0.0034\n",
      "Epoch 46/100\n",
      "97/97 [==============================] - 1s 9ms/step - loss: 0.0035\n",
      "Epoch 47/100\n",
      "97/97 [==============================] - 1s 8ms/step - loss: 0.0033\n",
      "Epoch 48/100\n",
      "97/97 [==============================] - 1s 6ms/step - loss: 0.0033\n",
      "Epoch 49/100\n",
      "97/97 [==============================] - 0s 5ms/step - loss: 0.0034\n",
      "Epoch 50/100\n",
      "97/97 [==============================] - 1s 8ms/step - loss: 0.0034\n",
      "Epoch 51/100\n",
      "97/97 [==============================] - 1s 8ms/step - loss: 0.0034\n",
      "Epoch 52/100\n",
      "97/97 [==============================] - 1s 6ms/step - loss: 0.0033\n",
      "Epoch 53/100\n",
      "97/97 [==============================] - 0s 5ms/step - loss: 0.0033\n",
      "Epoch 54/100\n",
      "97/97 [==============================] - 0s 5ms/step - loss: 0.0034\n",
      "Epoch 55/100\n",
      "97/97 [==============================] - 1s 9ms/step - loss: 0.0034\n",
      "Epoch 56/100\n",
      "97/97 [==============================] - 1s 5ms/step - loss: 0.0032\n",
      "Epoch 57/100\n",
      "97/97 [==============================] - 0s 5ms/step - loss: 0.0034\n",
      "Epoch 58/100\n",
      "97/97 [==============================] - 0s 5ms/step - loss: 0.0032\n",
      "Epoch 59/100\n",
      "97/97 [==============================] - 1s 5ms/step - loss: 0.0033\n",
      "Epoch 60/100\n",
      "97/97 [==============================] - 1s 6ms/step - loss: 0.0033\n",
      "Epoch 61/100\n",
      "97/97 [==============================] - 1s 6ms/step - loss: 0.0032\n",
      "Epoch 62/100\n",
      "97/97 [==============================] - 0s 5ms/step - loss: 0.0032\n",
      "Epoch 63/100\n",
      "97/97 [==============================] - 0s 5ms/step - loss: 0.0030\n",
      "Epoch 64/100\n",
      "97/97 [==============================] - 0s 5ms/step - loss: 0.0032\n",
      "Epoch 65/100\n",
      "97/97 [==============================] - 1s 7ms/step - loss: 0.0031\n",
      "Epoch 66/100\n",
      "97/97 [==============================] - 0s 4ms/step - loss: 0.0032\n",
      "Epoch 67/100\n",
      "97/97 [==============================] - 0s 4ms/step - loss: 0.0032\n",
      "Epoch 68/100\n",
      "97/97 [==============================] - 0s 4ms/step - loss: 0.0032\n",
      "Epoch 69/100\n",
      "97/97 [==============================] - 0s 4ms/step - loss: 0.0031\n",
      "Epoch 70/100\n",
      "97/97 [==============================] - 0s 4ms/step - loss: 0.0030\n",
      "Epoch 71/100\n",
      "97/97 [==============================] - 0s 4ms/step - loss: 0.0030\n",
      "Epoch 72/100\n",
      "97/97 [==============================] - 0s 4ms/step - loss: 0.0031\n",
      "Epoch 73/100\n",
      "97/97 [==============================] - 0s 4ms/step - loss: 0.0029\n",
      "Epoch 74/100\n",
      "97/97 [==============================] - 0s 4ms/step - loss: 0.0030\n",
      "Epoch 75/100\n",
      "97/97 [==============================] - 1s 8ms/step - loss: 0.0030\n",
      "Epoch 76/100\n",
      "97/97 [==============================] - 0s 4ms/step - loss: 0.0027\n",
      "Epoch 77/100\n",
      "97/97 [==============================] - 0s 5ms/step - loss: 0.0028\n",
      "Epoch 78/100\n",
      "97/97 [==============================] - 1s 6ms/step - loss: 0.0028\n",
      "Epoch 79/100\n",
      "97/97 [==============================] - 0s 4ms/step - loss: 0.0028\n",
      "Epoch 80/100\n",
      "97/97 [==============================] - 0s 4ms/step - loss: 0.0028\n",
      "Epoch 81/100\n",
      "97/97 [==============================] - 1s 8ms/step - loss: 0.0028\n",
      "Epoch 82/100\n",
      "97/97 [==============================] - 0s 4ms/step - loss: 0.0027\n",
      "Epoch 83/100\n",
      "97/97 [==============================] - 0s 4ms/step - loss: 0.0027\n",
      "Epoch 84/100\n",
      "97/97 [==============================] - 0s 4ms/step - loss: 0.0026\n",
      "Epoch 85/100\n",
      "97/97 [==============================] - 0s 4ms/step - loss: 0.0026\n",
      "Epoch 86/100\n",
      "97/97 [==============================] - 0s 5ms/step - loss: 0.0025\n",
      "Epoch 87/100\n",
      "97/97 [==============================] - 1s 6ms/step - loss: 0.0026\n",
      "Epoch 88/100\n",
      "97/97 [==============================] - 0s 4ms/step - loss: 0.0026\n",
      "Epoch 89/100\n",
      "97/97 [==============================] - 0s 4ms/step - loss: 0.0025\n",
      "Epoch 90/100\n",
      "97/97 [==============================] - 0s 4ms/step - loss: 0.0024\n",
      "Epoch 91/100\n",
      "97/97 [==============================] - 0s 4ms/step - loss: 0.0024\n",
      "Epoch 92/100\n",
      "97/97 [==============================] - 0s 4ms/step - loss: 0.0024\n",
      "Epoch 93/100\n",
      "97/97 [==============================] - 1s 7ms/step - loss: 0.0024\n",
      "Epoch 94/100\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "97/97 [==============================] - 1s 6ms/step - loss: 0.0024\n",
      "Epoch 95/100\n",
      "97/97 [==============================] - 1s 5ms/step - loss: 0.0023\n",
      "Epoch 96/100\n",
      "97/97 [==============================] - 0s 4ms/step - loss: 0.0023\n",
      "Epoch 97/100\n",
      "97/97 [==============================] - 0s 4ms/step - loss: 0.0023\n",
      "Epoch 98/100\n",
      "97/97 [==============================] - 1s 7ms/step - loss: 0.0023\n",
      "Epoch 99/100\n",
      "97/97 [==============================] - 0s 4ms/step - loss: 0.0023\n",
      "Epoch 100/100\n",
      "97/97 [==============================] - 1s 6ms/step - loss: 0.0022\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1c2a1baef0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "tf.set_random_seed(1)\n",
    "np.random.seed(1)\n",
    "\n",
    "sequence_len = 3\n",
    "X_train, y_train = create_sequences(training, sequence_len)\n",
    "X_test, y_test = create_sequences(testing, sequence_len)\n",
    "\n",
    "model = Sequential()\n",
    "model.add(LSTM(4, activation=\"tanh\", input_shape=(sequence_len, 1)))\n",
    "model.add(Dense(1))\n",
    "model.compile(loss='mean_squared_error', optimizer='adam')\n",
    "model.summary()\n",
    "\n",
    "tensor_board = TensorBoard(log_dir=\"/tmp/tf/logs/%d\" % time())\n",
    "model.fit(X_train.reshape(-1, sequence_len, 1), y_train\n",
    "          , epochs=100\n",
    "          , batch_size=1\n",
    "          , verbose=1\n",
    "          , callbacks=[tensor_board])\n",
    "\n",
    "y_train_pred = model.predict(X_train.reshape(-1, sequence_len, 1))\n",
    "y_test_pred = model.predict(X_test.reshape(-1, sequence_len, 1))\n",
    "\n",
    "show_values(y_train, y_test, y_train_pred, y_test_pred)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Stateful and stacked with time step and window"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "lstm_3 (LSTM)                (1, 3, 4)                 96        \n",
      "_________________________________________________________________\n",
      "dropout_1 (Dropout)          (1, 3, 4)                 0         \n",
      "_________________________________________________________________\n",
      "lstm_4 (LSTM)                (1, 4)                    144       \n",
      "_________________________________________________________________\n",
      "dense_3 (Dense)              (1, 1)                    5         \n",
      "=================================================================\n",
      "Total params: 245\n",
      "Trainable params: 245\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n",
      "Epoch 1/1\n",
      "97/97 [==============================] - 1s 10ms/step - loss: 0.0055\n",
      "Epoch 1/1\n",
      "97/97 [==============================] - 1s 5ms/step - loss: 0.0153\n",
      "Epoch 1/1\n",
      "97/97 [==============================] - 1s 6ms/step - loss: 0.0124\n",
      "Epoch 1/1\n",
      "97/97 [==============================] - 0s 5ms/step - loss: 0.0108\n",
      "Epoch 1/1\n",
      "97/97 [==============================] - 0s 5ms/step - loss: 0.0074\n",
      "Epoch 1/1\n",
      "97/97 [==============================] - 0s 5ms/step - loss: 0.0069\n",
      "Epoch 1/1\n",
      "97/97 [==============================] - 1s 6ms/step - loss: 0.0066\n",
      "Epoch 1/1\n",
      "97/97 [==============================] - 0s 5ms/step - loss: 0.0066\n",
      "Epoch 1/1\n",
      "97/97 [==============================] - 0s 5ms/step - loss: 0.0055\n",
      "Epoch 1/1\n",
      "97/97 [==============================] - 0s 5ms/step - loss: 0.0065\n",
      "Epoch 1/1\n",
      "97/97 [==============================] - 1s 5ms/step - loss: 0.0063\n",
      "Epoch 1/1\n",
      "97/97 [==============================] - 1s 8ms/step - loss: 0.0067\n",
      "Epoch 1/1\n",
      "97/97 [==============================] - 0s 5ms/step - loss: 0.0054\n",
      "Epoch 1/1\n",
      "97/97 [==============================] - 0s 5ms/step - loss: 0.0055\n",
      "Epoch 1/1\n",
      "97/97 [==============================] - 0s 5ms/step - loss: 0.0062\n",
      "Epoch 1/1\n",
      "97/97 [==============================] - 0s 5ms/step - loss: 0.0052\n",
      "Epoch 1/1\n",
      "97/97 [==============================] - 0s 5ms/step - loss: 0.0069\n",
      "Epoch 1/1\n",
      "97/97 [==============================] - 1s 5ms/step - loss: 0.0071\n",
      "Epoch 1/1\n",
      "97/97 [==============================] - 1s 7ms/step - loss: 0.0055\n",
      "Epoch 1/1\n",
      "97/97 [==============================] - 0s 5ms/step - loss: 0.0057\n",
      "Epoch 1/1\n",
      "97/97 [==============================] - 0s 5ms/step - loss: 0.0056\n",
      "Epoch 1/1\n",
      "97/97 [==============================] - 1s 7ms/step - loss: 0.0056\n",
      "Epoch 1/1\n",
      "97/97 [==============================] - 0s 5ms/step - loss: 0.0052\n",
      "Epoch 1/1\n",
      "97/97 [==============================] - 0s 5ms/step - loss: 0.0058\n",
      "Epoch 1/1\n",
      "97/97 [==============================] - 1s 7ms/step - loss: 0.0055\n",
      "Epoch 1/1\n",
      "97/97 [==============================] - 1s 7ms/step - loss: 0.0059\n",
      "Epoch 1/1\n",
      "97/97 [==============================] - 1s 7ms/step - loss: 0.0049\n",
      "Epoch 1/1\n",
      "97/97 [==============================] - 1s 7ms/step - loss: 0.0067\n",
      "Epoch 1/1\n",
      "97/97 [==============================] - 1s 6ms/step - loss: 0.0066\n",
      "Epoch 1/1\n",
      "97/97 [==============================] - 1s 6ms/step - loss: 0.0061\n",
      "Epoch 1/1\n",
      "97/97 [==============================] - 1s 5ms/step - loss: 0.0056\n",
      "Epoch 1/1\n",
      "97/97 [==============================] - 1s 6ms/step - loss: 0.0051\n",
      "Epoch 1/1\n",
      "97/97 [==============================] - 1s 6ms/step - loss: 0.0059\n",
      "Epoch 1/1\n",
      "97/97 [==============================] - 1s 5ms/step - loss: 0.0062\n",
      "Epoch 1/1\n",
      "97/97 [==============================] - 1s 7ms/step - loss: 0.0063\n",
      "Epoch 1/1\n",
      "97/97 [==============================] - 1s 5ms/step - loss: 0.0058\n",
      "Epoch 1/1\n",
      "97/97 [==============================] - 1s 5ms/step - loss: 0.0063\n",
      "Epoch 1/1\n",
      "97/97 [==============================] - 1s 5ms/step - loss: 0.0058\n",
      "Epoch 1/1\n",
      "97/97 [==============================] - 1s 7ms/step - loss: 0.0058\n",
      "Epoch 1/1\n",
      "97/97 [==============================] - 1s 7ms/step - loss: 0.0055\n",
      "Epoch 1/1\n",
      "97/97 [==============================] - 1s 7ms/step - loss: 0.0057\n",
      "Epoch 1/1\n",
      "97/97 [==============================] - 1s 5ms/step - loss: 0.0060\n",
      "Epoch 1/1\n",
      "97/97 [==============================] - 1s 6ms/step - loss: 0.0050\n",
      "Epoch 1/1\n",
      "97/97 [==============================] - 1s 5ms/step - loss: 0.0052\n",
      "Epoch 1/1\n",
      "97/97 [==============================] - 1s 7ms/step - loss: 0.0054\n",
      "Epoch 1/1\n",
      "97/97 [==============================] - 1s 6ms/step - loss: 0.0058\n",
      "Epoch 1/1\n",
      "97/97 [==============================] - 1s 6ms/step - loss: 0.0056\n",
      "Epoch 1/1\n",
      "97/97 [==============================] - 1s 7ms/step - loss: 0.0058\n",
      "Epoch 1/1\n",
      "97/97 [==============================] - 1s 5ms/step - loss: 0.0048\n",
      "Epoch 1/1\n",
      "97/97 [==============================] - 1s 6ms/step - loss: 0.0062\n",
      "Epoch 1/1\n",
      "97/97 [==============================] - 1s 5ms/step - loss: 0.0055\n",
      "Epoch 1/1\n",
      "97/97 [==============================] - 1s 6ms/step - loss: 0.0059\n",
      "Epoch 1/1\n",
      "97/97 [==============================] - 1s 5ms/step - loss: 0.0054\n",
      "Epoch 1/1\n",
      "97/97 [==============================] - 1s 6ms/step - loss: 0.0056\n",
      "Epoch 1/1\n",
      "97/97 [==============================] - 1s 6ms/step - loss: 0.0053\n",
      "Epoch 1/1\n",
      "97/97 [==============================] - 1s 5ms/step - loss: 0.0052\n",
      "Epoch 1/1\n",
      "97/97 [==============================] - 1s 5ms/step - loss: 0.0055\n",
      "Epoch 1/1\n",
      "97/97 [==============================] - 1s 6ms/step - loss: 0.0054\n",
      "Epoch 1/1\n",
      "97/97 [==============================] - 1s 5ms/step - loss: 0.0054\n",
      "Epoch 1/1\n",
      "97/97 [==============================] - 1s 5ms/step - loss: 0.0059\n",
      "Epoch 1/1\n",
      "97/97 [==============================] - 1s 6ms/step - loss: 0.0058\n",
      "Epoch 1/1\n",
      "97/97 [==============================] - 1s 5ms/step - loss: 0.0055\n",
      "Epoch 1/1\n",
      "97/97 [==============================] - 1s 6ms/step - loss: 0.0052\n",
      "Epoch 1/1\n",
      "97/97 [==============================] - 1s 5ms/step - loss: 0.0056\n",
      "Epoch 1/1\n",
      "97/97 [==============================] - 1s 7ms/step - loss: 0.0053\n",
      "Epoch 1/1\n",
      "97/97 [==============================] - 1s 6ms/step - loss: 0.0053\n",
      "Epoch 1/1\n",
      "97/97 [==============================] - 1s 5ms/step - loss: 0.0052\n",
      "Epoch 1/1\n",
      "97/97 [==============================] - 1s 8ms/step - loss: 0.0063\n",
      "Epoch 1/1\n",
      "97/97 [==============================] - 1s 5ms/step - loss: 0.0059\n",
      "Epoch 1/1\n",
      "97/97 [==============================] - 1s 6ms/step - loss: 0.0055\n",
      "Epoch 1/1\n",
      "97/97 [==============================] - 1s 6ms/step - loss: 0.0050\n",
      "Epoch 1/1\n",
      "97/97 [==============================] - 1s 5ms/step - loss: 0.0060\n",
      "Epoch 1/1\n",
      "97/97 [==============================] - 1s 7ms/step - loss: 0.0054\n",
      "Epoch 1/1\n",
      "97/97 [==============================] - 1s 5ms/step - loss: 0.0052\n",
      "Epoch 1/1\n",
      "97/97 [==============================] - 1s 6ms/step - loss: 0.0053\n",
      "Epoch 1/1\n",
      "97/97 [==============================] - 1s 5ms/step - loss: 0.0050\n",
      "Epoch 1/1\n",
      "97/97 [==============================] - 1s 5ms/step - loss: 0.0055\n",
      "Epoch 1/1\n",
      "97/97 [==============================] - 1s 6ms/step - loss: 0.0053\n",
      "Epoch 1/1\n",
      "97/97 [==============================] - 1s 6ms/step - loss: 0.0052\n",
      "Epoch 1/1\n",
      "97/97 [==============================] - 1s 5ms/step - loss: 0.0055\n",
      "Epoch 1/1\n",
      "97/97 [==============================] - 1s 5ms/step - loss: 0.0047\n",
      "Epoch 1/1\n",
      "97/97 [==============================] - 1s 5ms/step - loss: 0.0051\n",
      "Epoch 1/1\n",
      "97/97 [==============================] - 1s 5ms/step - loss: 0.0048\n",
      "Epoch 1/1\n",
      "97/97 [==============================] - 1s 8ms/step - loss: 0.0056\n",
      "Epoch 1/1\n",
      "97/97 [==============================] - 1s 6ms/step - loss: 0.0051\n",
      "Epoch 1/1\n",
      "97/97 [==============================] - 1s 7ms/step - loss: 0.0054\n",
      "Epoch 1/1\n",
      "97/97 [==============================] - 1s 7ms/step - loss: 0.0055\n",
      "Epoch 1/1\n",
      "97/97 [==============================] - 1s 7ms/step - loss: 0.0049\n",
      "Epoch 1/1\n",
      "97/97 [==============================] - 1s 5ms/step - loss: 0.0051\n",
      "Epoch 1/1\n",
      "97/97 [==============================] - 1s 5ms/step - loss: 0.0044\n",
      "Epoch 1/1\n",
      "97/97 [==============================] - 1s 6ms/step - loss: 0.0048\n",
      "Epoch 1/1\n",
      "97/97 [==============================] - 1s 5ms/step - loss: 0.0045\n",
      "Epoch 1/1\n",
      "97/97 [==============================] - 1s 7ms/step - loss: 0.0045\n",
      "Epoch 1/1\n",
      "97/97 [==============================] - 1s 5ms/step - loss: 0.0046\n",
      "Epoch 1/1\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "97/97 [==============================] - 1s 5ms/step - loss: 0.0046\n",
      "Epoch 1/1\n",
      "97/97 [==============================] - 1s 6ms/step - loss: 0.0059\n",
      "Epoch 1/1\n",
      "97/97 [==============================] - 1s 5ms/step - loss: 0.0047\n",
      "Epoch 1/1\n",
      "97/97 [==============================] - 1s 5ms/step - loss: 0.0049\n",
      "Epoch 1/1\n",
      "97/97 [==============================] - 1s 5ms/step - loss: 0.0048\n",
      "Epoch 1/1\n",
      "97/97 [==============================] - 1s 6ms/step - loss: 0.0046\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1c2a1baf98>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "tf.set_random_seed(1)\n",
    "np.random.seed(1)\n",
    "\n",
    "sequence_len = 3\n",
    "X_train, y_train = create_sequences(training, sequence_len)\n",
    "X_test, y_test = create_sequences(testing, sequence_len)\n",
    "\n",
    "model = Sequential()\n",
    "model.add(LSTM(4, batch_input_shape = (1, sequence_len, 1)\n",
    "               , input_shape=(sequence_len, 1), stateful=True, return_sequences=True))\n",
    "model.add(Dropout(rate=0.2))\n",
    "model.add(LSTM(4, input_shape=(sequence_len, 1)))\n",
    "model.add(Dense(1, activation = None))\n",
    "model.compile(loss='mean_squared_error', optimizer='adam')\n",
    "model.summary()\n",
    "\n",
    "tensor_board = TensorBoard(log_dir=\"/tmp/tf/logs/%d\" % time())\n",
    "for i in range(100):\n",
    "    model.fit(X_train.reshape(-1, sequence_len, 1), y_train \n",
    "              , epochs=1\n",
    "              , batch_size=1\n",
    "              , verbose=1\n",
    "              , shuffle=False\n",
    "              , callbacks=[tensor_board])\n",
    "    model.reset_states()\n",
    "\n",
    "y_train_pred = model.predict(X_train.reshape(-1, sequence_len, 1), batch_size=1)\n",
    "y_test_pred = model.predict(X_test.reshape(-1, sequence_len, 1), batch_size=1)\n",
    "\n",
    "show_values(y_train, y_test, y_train_pred, y_test_pred)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "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": "3.6.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}