models/failure_prediction/jnotebooks/LSTM.ipynb

   1 {
   2   "nbformat": 4,
   3   "nbformat_minor": 0,
   4   "metadata": {
   5     "colab": {
   6       "name": "LSTM.ipynb",
   7       "provenance": [],
   8       "collapsed_sections": []
   9     },
  10     "kernelspec": {
  11       "display_name": "Python 3",
  12       "name": "python3"
  13     },
  14     "language_info": {
  15       "name": "python"
  16     }
  17   },
  18   "cells": [
  19     {
  20       "cell_type": "markdown",
  21       "metadata": {
  22         "id": "mF-r4kHqhhr0"
  23       },
  24       "source": [
  25         "Contributors: Rohit Singh Rathaur, Girish L.\n",
  26         "\n",
  27         "Copyright 2021 [Rohit Singh Rathaur, BIT Mesra and Girish L., CIT GUBBI, Karnataka]\n",
  28         "\n",
  29         "Licensed under the Apache License, Version 2.0 (the \"License\"); you may not use this file except in compliance with the License. You may obtain a copy of the License at\n",
  30         "\n",
  31         "http://www.apache.org/licenses/LICENSE-2.0\n",
  32         "Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on an \"AS IS\" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the specific language governing permissions and limitations under the License."
  33       ]
  34     },
  35     {
  36       "cell_type": "markdown",
  37       "metadata": {
  38         "id": "0lwwYh5Yhk-n"
  39       },
  40       "source": [
  41         "We mounted the drive to access the data from google drive"
  42       ]
  43     },
  44     {
  45       "cell_type": "code",
  46       "metadata": {
  47         "colab": {
  48           "base_uri": "https://localhost:8080/"
  49         },
  50         "id": "YQ6lT1e2hrx4",
  51         "outputId": "552d1331-3807-437c-aded-2e495ae0f860"
  52       },
  53       "source": [
  54         "from google.colab import drive\n",
  55         "drive.mount('/content/drive')"
  56       ],
  57       "execution_count": null,
  58       "outputs": [
  59         {
  60           "output_type": "stream",
  61           "name": "stdout",
  62           "text": [
  63             "Drive already mounted at /content/drive; to attempt to forcibly remount, call drive.mount(\"/content/drive\", force_remount=True).\n"
  64           ]
  65         }
  66       ]
  67     },
  68     {
  69       "cell_type": "markdown",
  70       "metadata": {
  71         "id": "OwlM7YM2htFR"
  72       },
  73       "source": [
  74         "We are importing the libraries:\n",
  75         "- TensorFlow: to process and train the model\n",
  76         "- Matplotlib: to plot the training anf loss curves\n",
  77         "- Pandas: used for data analysis and it allows us to import data from various formats\n",
  78         "- Numpy: For array computing"
  79       ]
  80     },
  81     {
  82       "cell_type": "code",
  83       "metadata": {
  84         "id": "tLhroy5BnMnC"
  85       },
  86       "source": [
  87         "# Importing libraries\n",
  88         "import tensorflow as tf\n",
  89         "import matplotlib.pyplot as plt\n",
  90         "import matplotlib as mpl\n",
  91         "import pandas as pd\n",
  92         "import numpy as np\n",
  93         "import os"
  94       ],
  95       "execution_count": null,
  96       "outputs": []
  97     },
  98     {
  99       "cell_type": "markdown",
 100       "metadata": {
 101         "id": "qlVZPlu-iiYw"
 102       },
 103       "source": [
 104         "We are reading the CSV file using `read_csv` function and storing it in a DataFrame named `df_Ellis`"
 105       ]
 106     },
 107     {
 108       "cell_type": "code",
 109       "metadata": {
 110         "colab": {
 111           "base_uri": "https://localhost:8080/",
 112           "height": 419
 113         },
 114         "id": "2-UpMVsSnfCI",
 115         "outputId": "06698528-eedb-475b-fa32-e347d03e9980"
 116       },
 117       "source": [
 118         "df_Ellis  = pd.read_csv(\"/content/drive/MyDrive/Failure/lstm/Ellis_FinalTwoConditionwithOR.csv\")\n",
 119         "df_Ellis"
 120       ],
 121       "execution_count": null,
 122       "outputs": [
 123         {
 124           "output_type": "execute_result",
 125           "data": {
 126             "text/html": [
 127               "<div>\n",
 128               "<style scoped>\n",
 129               "    .dataframe tbody tr th:only-of-type {\n",
 130               "        vertical-align: middle;\n",
 131               "    }\n",
 132               "\n",
 133               "    .dataframe tbody tr th {\n",
 134               "        vertical-align: top;\n",
 135               "    }\n",
 136               "\n",
 137               "    .dataframe thead th {\n",
 138               "        text-align: right;\n",
 139               "    }\n",
 140               "</style>\n",
 141               "<table border=\"1\" class=\"dataframe\">\n",
 142               "  <thead>\n",
 143               "    <tr style=\"text-align: right;\">\n",
 144               "      <th></th>\n",
 145               "      <th>Timestamp</th>\n",
 146               "      <th>ellis-cpu.system_perc</th>\n",
 147               "      <th>ellis-cpu.wait_perc</th>\n",
 148               "      <th>ellis-load.avg_1_min</th>\n",
 149               "      <th>ellis-mem.free_mb</th>\n",
 150               "      <th>ellis-net.in_bytes_sec</th>\n",
 151               "      <th>ellis-net.out_packets_sec</th>\n",
 152               "      <th>Label</th>\n",
 153               "    </tr>\n",
 154               "  </thead>\n",
 155               "  <tbody>\n",
 156               "    <tr>\n",
 157               "      <th>0</th>\n",
 158               "      <td>14-09-2016 0:00</td>\n",
 159               "      <td>0.5</td>\n",
 160               "      <td>12.9</td>\n",
 161               "      <td>1.730</td>\n",
 162               "      <td>3949</td>\n",
 163               "      <td>5413.200</td>\n",
 164               "      <td>62.067</td>\n",
 165               "      <td>1</td>\n",
 166               "    </tr>\n",
 167               "    <tr>\n",
 168               "      <th>1</th>\n",
 169               "      <td>14-09-2016 0:00</td>\n",
 170               "      <td>0.4</td>\n",
 171               "      <td>10.3</td>\n",
 172               "      <td>1.790</td>\n",
 173               "      <td>3950</td>\n",
 174               "      <td>5201.667</td>\n",
 175               "      <td>59.567</td>\n",
 176               "      <td>1</td>\n",
 177               "    </tr>\n",
 178               "    <tr>\n",
 179               "      <th>2</th>\n",
 180               "      <td>14-09-2016 0:01</td>\n",
 181               "      <td>0.4</td>\n",
 182               "      <td>11.8</td>\n",
 183               "      <td>1.520</td>\n",
 184               "      <td>3950</td>\n",
 185               "      <td>5370.733</td>\n",
 186               "      <td>61.200</td>\n",
 187               "      <td>1</td>\n",
 188               "    </tr>\n",
 189               "    <tr>\n",
 190               "      <th>3</th>\n",
 191               "      <td>14-09-2016 0:01</td>\n",
 192               "      <td>0.4</td>\n",
 193               "      <td>12.9</td>\n",
 194               "      <td>1.430</td>\n",
 195               "      <td>3949</td>\n",
 196               "      <td>5292.467</td>\n",
 197               "      <td>60.400</td>\n",
 198               "      <td>1</td>\n",
 199               "    </tr>\n",
 200               "    <tr>\n",
 201               "      <th>4</th>\n",
 202               "      <td>14-09-2016 0:02</td>\n",
 203               "      <td>0.5</td>\n",
 204               "      <td>12.1</td>\n",
 205               "      <td>1.440</td>\n",
 206               "      <td>3950</td>\n",
 207               "      <td>5318.167</td>\n",
 208               "      <td>61.700</td>\n",
 209               "      <td>1</td>\n",
 210               "    </tr>\n",
 211               "    <tr>\n",
 212               "      <th>...</th>\n",
 213               "      <td>...</td>\n",
 214               "      <td>...</td>\n",
 215               "      <td>...</td>\n",
 216               "      <td>...</td>\n",
 217               "      <td>...</td>\n",
 218               "      <td>...</td>\n",
 219               "      <td>...</td>\n",
 220               "      <td>...</td>\n",
 221               "    </tr>\n",
 222               "    <tr>\n",
 223               "      <th>176995</th>\n",
 224               "      <td>13-12-2016 21:20</td>\n",
 225               "      <td>0.4</td>\n",
 226               "      <td>0.3</td>\n",
 227               "      <td>0.030</td>\n",
 228               "      <td>3484</td>\n",
 229               "      <td>230.967</td>\n",
 230               "      <td>2.167</td>\n",
 231               "      <td>0</td>\n",
 232               "    </tr>\n",
 233               "    <tr>\n",
 234               "      <th>176996</th>\n",
 235               "      <td>13-12-2016 21:20</td>\n",
 236               "      <td>0.2</td>\n",
 237               "      <td>0.3</td>\n",
 238               "      <td>0.018</td>\n",
 239               "      <td>3484</td>\n",
 240               "      <td>218.433</td>\n",
 241               "      <td>0.767</td>\n",
 242               "      <td>0</td>\n",
 243               "    </tr>\n",
 244               "    <tr>\n",
 245               "      <th>176997</th>\n",
 246               "      <td>13-12-2016 21:21</td>\n",
 247               "      <td>0.6</td>\n",
 248               "      <td>0.3</td>\n",
 249               "      <td>0.010</td>\n",
 250               "      <td>3483</td>\n",
 251               "      <td>160.967</td>\n",
 252               "      <td>1.867</td>\n",
 253               "      <td>0</td>\n",
 254               "    </tr>\n",
 255               "    <tr>\n",
 256               "      <th>176998</th>\n",
 257               "      <td>13-12-2016 21:21</td>\n",
 258               "      <td>0.6</td>\n",
 259               "      <td>0.3</td>\n",
 260               "      <td>0.007</td>\n",
 261               "      <td>3484</td>\n",
 262               "      <td>188.367</td>\n",
 263               "      <td>2.100</td>\n",
 264               "      <td>0</td>\n",
 265               "    </tr>\n",
 266               "    <tr>\n",
 267               "      <th>176999</th>\n",
 268               "      <td>13-12-2016 21:22</td>\n",
 269               "      <td>0.4</td>\n",
 270               "      <td>0.1</td>\n",
 271               "      <td>0.040</td>\n",
 272               "      <td>3484</td>\n",
 273               "      <td>229.833</td>\n",
 274               "      <td>2.400</td>\n",
 275               "      <td>0</td>\n",
 276               "    </tr>\n",
 277               "  </tbody>\n",
 278               "</table>\n",
 279               "<p>177000 rows × 8 columns</p>\n",
 280               "</div>"
 281             ],
 282             "text/plain": [
 283               "               Timestamp  ...  Label\n",
 284               "0        14-09-2016 0:00  ...      1\n",
 285               "1        14-09-2016 0:00  ...      1\n",
 286               "2        14-09-2016 0:01  ...      1\n",
 287               "3        14-09-2016 0:01  ...      1\n",
 288               "4        14-09-2016 0:02  ...      1\n",
 289               "...                  ...  ...    ...\n",
 290               "176995  13-12-2016 21:20  ...      0\n",
 291               "176996  13-12-2016 21:20  ...      0\n",
 292               "176997  13-12-2016 21:21  ...      0\n",
 293               "176998  13-12-2016 21:21  ...      0\n",
 294               "176999  13-12-2016 21:22  ...      0\n",
 295               "\n",
 296               "[177000 rows x 8 columns]"
 297             ]
 298           },
 299           "metadata": {},
 300           "execution_count": 7
 301         }
 302       ]
 303     },
 304     {
 305       "cell_type": "markdown",
 306       "metadata": {
 307         "id": "t5gl1fqmjBdt"
 308       },
 309       "source": [
 310         "`plot()` function is used to draw points"
 311       ]
 312     },
 313     {
 314       "cell_type": "code",
 315       "metadata": {
 316         "colab": {
 317           "base_uri": "https://localhost:8080/",
 318           "height": 293
 319         },
 320         "id": "92xBt43BnjAo",
 321         "outputId": "547a5311-6db3-42b6-d5bd-3bc171446eac"
 322       },
 323       "source": [
 324         "df_Ellis.plot()"
 325       ],
 326       "execution_count": null,
 327       "outputs": [
 328         {
 329           "output_type": "execute_result",
 330           "data": {
 331             "text/plain": [
 332               "<matplotlib.axes._subplots.AxesSubplot at 0x7fbb837d7f90>"
 333             ]
 334           },
 335           "metadata": {},
 336           "execution_count": 8
 337         },
 338         {
 339           "output_type": "display_data",
 340           "data": {
 341             "image/png": "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\n",
 342             "text/plain": [
 343               "<Figure size 432x288 with 1 Axes>"
 344             ]
 345           },
 346           "metadata": {
 347             "needs_background": "light"
 348           }
 349         }
 350       ]
 351     },
 352     {
 353       "cell_type": "markdown",
 354       "metadata": {
 355         "id": "KYRyARNcjJkz"
 356       },
 357       "source": [
 358         "Using multivariate features:\n",
 359         "- Storing only the multivariate features in a dataframe named `features_3`\n",
 360         "- Extracting the Timestamp column from `df_Ellis` dataframe\n",
 361         "- and combining it with the dataframe `features`"
 362       ]
 363     },
 364     {
 365       "cell_type": "code",
 366       "metadata": {
 367         "colab": {
 368           "base_uri": "https://localhost:8080/",
 369           "height": 235
 370         },
 371         "id": "5oQK-ddinvCM",
 372         "outputId": "8bb3a951-1f2d-4f40-aa40-668e00bdc238"
 373       },
 374       "source": [
 375         "## ## using multivariate feature \n",
 376         "\n",
 377         "features_3 = ['ellis-cpu.system_perc', 'ellis-cpu.wait_perc', 'ellis-load.avg_1_min', 'ellis-mem.free_mb', 'ellis-net.in_bytes_sec', 'ellis-net.out_packets_sec', 'Label']\n",
 378         "\n",
 379         "features = df_Ellis[features_3]\n",
 380         "features.index = df_Ellis['Timestamp']\n",
 381         "features.head()"
 382       ],
 383       "execution_count": null,
 384       "outputs": [
 385         {
 386           "output_type": "execute_result",
 387           "data": {
 388             "text/html": [
 389               "<div>\n",
 390               "<style scoped>\n",
 391               "    .dataframe tbody tr th:only-of-type {\n",
 392               "        vertical-align: middle;\n",
 393               "    }\n",
 394               "\n",
 395               "    .dataframe tbody tr th {\n",
 396               "        vertical-align: top;\n",
 397               "    }\n",
 398               "\n",
 399               "    .dataframe thead th {\n",
 400               "        text-align: right;\n",
 401               "    }\n",
 402               "</style>\n",
 403               "<table border=\"1\" class=\"dataframe\">\n",
 404               "  <thead>\n",
 405               "    <tr style=\"text-align: right;\">\n",
 406               "      <th></th>\n",
 407               "      <th>ellis-cpu.system_perc</th>\n",
 408               "      <th>ellis-cpu.wait_perc</th>\n",
 409               "      <th>ellis-load.avg_1_min</th>\n",
 410               "      <th>ellis-mem.free_mb</th>\n",
 411               "      <th>ellis-net.in_bytes_sec</th>\n",
 412               "      <th>ellis-net.out_packets_sec</th>\n",
 413               "      <th>Label</th>\n",
 414               "    </tr>\n",
 415               "    <tr>\n",
 416               "      <th>Timestamp</th>\n",
 417               "      <th></th>\n",
 418               "      <th></th>\n",
 419               "      <th></th>\n",
 420               "      <th></th>\n",
 421               "      <th></th>\n",
 422               "      <th></th>\n",
 423               "      <th></th>\n",
 424               "    </tr>\n",
 425               "  </thead>\n",
 426               "  <tbody>\n",
 427               "    <tr>\n",
 428               "      <th>14-09-2016 0:00</th>\n",
 429               "      <td>0.5</td>\n",
 430               "      <td>12.9</td>\n",
 431               "      <td>1.73</td>\n",
 432               "      <td>3949</td>\n",
 433               "      <td>5413.200</td>\n",
 434               "      <td>62.067</td>\n",
 435               "      <td>1</td>\n",
 436               "    </tr>\n",
 437               "    <tr>\n",
 438               "      <th>14-09-2016 0:00</th>\n",
 439               "      <td>0.4</td>\n",
 440               "      <td>10.3</td>\n",
 441               "      <td>1.79</td>\n",
 442               "      <td>3950</td>\n",
 443               "      <td>5201.667</td>\n",
 444               "      <td>59.567</td>\n",
 445               "      <td>1</td>\n",
 446               "    </tr>\n",
 447               "    <tr>\n",
 448               "      <th>14-09-2016 0:01</th>\n",
 449               "      <td>0.4</td>\n",
 450               "      <td>11.8</td>\n",
 451               "      <td>1.52</td>\n",
 452               "      <td>3950</td>\n",
 453               "      <td>5370.733</td>\n",
 454               "      <td>61.200</td>\n",
 455               "      <td>1</td>\n",
 456               "    </tr>\n",
 457               "    <tr>\n",
 458               "      <th>14-09-2016 0:01</th>\n",
 459               "      <td>0.4</td>\n",
 460               "      <td>12.9</td>\n",
 461               "      <td>1.43</td>\n",
 462               "      <td>3949</td>\n",
 463               "      <td>5292.467</td>\n",
 464               "      <td>60.400</td>\n",
 465               "      <td>1</td>\n",
 466               "    </tr>\n",
 467               "    <tr>\n",
 468               "      <th>14-09-2016 0:02</th>\n",
 469               "      <td>0.5</td>\n",
 470               "      <td>12.1</td>\n",
 471               "      <td>1.44</td>\n",
 472               "      <td>3950</td>\n",
 473               "      <td>5318.167</td>\n",
 474               "      <td>61.700</td>\n",
 475               "      <td>1</td>\n",
 476               "    </tr>\n",
 477               "  </tbody>\n",
 478               "</table>\n",
 479               "</div>"
 480             ],
 481             "text/plain": [
 482               "                 ellis-cpu.system_perc  ...  Label\n",
 483               "Timestamp                               ...       \n",
 484               "14-09-2016 0:00                    0.5  ...      1\n",
 485               "14-09-2016 0:00                    0.4  ...      1\n",
 486               "14-09-2016 0:01                    0.4  ...      1\n",
 487               "14-09-2016 0:01                    0.4  ...      1\n",
 488               "14-09-2016 0:02                    0.5  ...      1\n",
 489               "\n",
 490               "[5 rows x 7 columns]"
 491             ]
 492           },
 493           "metadata": {},
 494           "execution_count": 9
 495         }
 496       ]
 497     },
 498     {
 499       "cell_type": "markdown",
 500       "metadata": {
 501         "id": "jO-JgNPQj99-"
 502       },
 503       "source": [
 504         "Plotted features"
 505       ]
 506     },
 507     {
 508       "cell_type": "code",
 509       "metadata": {
 510         "colab": {
 511           "base_uri": "https://localhost:8080/",
 512           "height": 437
 513         },
 514         "id": "qbqn755fo81g",
 515         "outputId": "3a6ee243-5136-4614-93e6-8c2f7f4c13af"
 516       },
 517       "source": [
 518         "features.plot(subplots=True)"
 519       ],
 520       "execution_count": null,
 521       "outputs": [
 522         {
 523           "output_type": "execute_result",
 524           "data": {
 525             "text/plain": [
 526               "array([<matplotlib.axes._subplots.AxesSubplot object at 0x7fbb836b5d10>,\n",
 527               "       <matplotlib.axes._subplots.AxesSubplot object at 0x7fbb83221150>,\n",
 528               "       <matplotlib.axes._subplots.AxesSubplot object at 0x7fbb836b5690>,\n",
 529               "       <matplotlib.axes._subplots.AxesSubplot object at 0x7fbb8318f890>,\n",
 530               "       <matplotlib.axes._subplots.AxesSubplot object at 0x7fbb83146c50>,\n",
 531               "       <matplotlib.axes._subplots.AxesSubplot object at 0x7fbb830fefd0>,\n",
 532               "       <matplotlib.axes._subplots.AxesSubplot object at 0x7fbb830c3490>],\n",
 533               "      dtype=object)"
 534             ]
 535           },
 536           "metadata": {},
 537           "execution_count": 10
 538         },
 539         {
 540           "output_type": "display_data",
 541           "data": {
 542             "image/png": "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\n",
 543             "text/plain": [
 544               "<Figure size 432x288 with 7 Axes>"
 545             ]
 546           },
 547           "metadata": {
 548             "needs_background": "light"
 549           }
 550         }
 551       ]
 552     },
 553     {
 554       "cell_type": "code",
 555       "metadata": {
 556         "id": "jJQD1x9psWCH"
 557       },
 558       "source": [
 559         "features = features.values"
 560       ],
 561       "execution_count": null,
 562       "outputs": []
 563     },
 564     {
 565       "cell_type": "code",
 566       "metadata": {
 567         "colab": {
 568           "base_uri": "https://localhost:8080/"
 569         },
 570         "id": "xf8WCiykpUzN",
 571         "outputId": "9337b00a-1d3f-4dbc-91d5-2f445eda21e5"
 572       },
 573       "source": [
 574         "### standardize data\n",
 575         "train_split = 141600\n",
 576         "tf.random.set_seed(13)\n",
 577         "\n",
 578         "### standardize data\n",
 579         "features_mean = features[:train_split].mean()\n",
 580         "features_std = features[:train_split].std()\n",
 581         "features  = (features - features_mean)/ features_std\n",
 582         "\n",
 583         "print(type(features))\n",
 584         "print(features.shape)\n"
 585       ],
 586       "execution_count": null,
 587       "outputs": [
 588         {
 589           "output_type": "stream",
 590           "name": "stdout",
 591           "text": [
 592             "<class 'numpy.ndarray'>\n",
 593             "(177000, 7)\n"
 594           ]
 595         }
 596       ]
 597     },
 598     {
 599       "cell_type": "code",
 600       "metadata": {
 601         "id": "1a0hNDmppnLB"
 602       },
 603       "source": [
 604         "### create mutlivariate data\n",
 605         "\n",
 606         "def mutlivariate_data(features , target , start_idx , end_idx , history_size , target_size,\n",
 607         "                      step ,  single_step = False):\n",
 608         "  data = []\n",
 609         "  labels = []\n",
 610         "  start_idx = start_idx + history_size\n",
 611         "  if end_idx is None:\n",
 612         "    end_idx = len(features)- target_size\n",
 613         "  for i in range(start_idx , end_idx ):\n",
 614         "    idxs = range(i-history_size, i, step) ### using step\n",
 615         "    data.append(features[idxs])\n",
 616         "    if single_step:\n",
 617         "      labels.append(target[i+target_size])\n",
 618         "    else:\n",
 619         "      labels.append(target[i:i+target_size])\n",
 620         "\n",
 621         "  return np.array(data) , np.array(labels)"
 622       ],
 623       "execution_count": null,
 624       "outputs": []
 625     },
 626     {
 627       "cell_type": "markdown",
 628       "metadata": {
 629         "id": "vOzDoRCrlTBc"
 630       },
 631       "source": [
 632         "We spliited the multivariate data in tarining and validation and printed the shape of that data."
 633       ]
 634     },
 635     {
 636       "cell_type": "code",
 637       "metadata": {
 638         "colab": {
 639           "base_uri": "https://localhost:8080/"
 640         },
 641         "id": "Z0CivgkitfgE",
 642         "outputId": "73165ad2-6043-42c9-ed7d-ff050af74e34"
 643       },
 644       "source": [
 645         "### generate multivariate data\n",
 646         "\n",
 647         "history = 720\n",
 648         "future_target = 72\n",
 649         "STEP = 6\n",
 650         "\n",
 651         "x_train_ss , y_train_ss = mutlivariate_data(features , features[:, 1], 0, train_split, history,\n",
 652         "                                            future_target, STEP , single_step = True)\n",
 653         "\n",
 654         "x_val_ss , y_val_ss = mutlivariate_data(features , features[:,1] , train_split , None , history ,\n",
 655         "                                        future_target, STEP, single_step = True)\n",
 656         "\n",
 657         "print(x_train_ss.shape , y_train_ss.shape)\n",
 658         "print(x_val_ss.shape , y_val_ss.shape)"
 659       ],
 660       "execution_count": null,
 661       "outputs": [
 662         {
 663           "output_type": "stream",
 664           "name": "stdout",
 665           "text": [
 666             "(140880, 120, 7) (140880,)\n",
 667             "(34608, 120, 7) (34608,)\n"
 668           ]
 669         }
 670       ]
 671     },
 672     {
 673       "cell_type": "markdown",
 674       "metadata": {
 675         "id": "YU_j8yNblhOO"
 676       },
 677       "source": [
 678         ""
 679       ]
 680     },
 681     {
 682       "cell_type": "markdown",
 683       "metadata": {
 684         "id": "QialFchyl9Rt"
 685       },
 686       "source": [
 687         "The `tf.data.Dataset` API supports writing descriptive and efficient input pipelines. Dataset usage following a common pattern:\n",
 688         "- Creating a source dataset from our input data.\n",
 689         "- Applied dataset transformations to preprocess the data.\n",
 690         "- Iterate over the dataset and process the elements.\n",
 691         "Note: Iteration happens in a streaming fashion, so the full dataset does not need to fit into memory.\n",
 692         "Once we have a dataset, we can apply transformations to prepare the data for our model:"
 693       ]
 694     },
 695     {
 696       "cell_type": "code",
 697       "metadata": {
 698         "colab": {
 699           "base_uri": "https://localhost:8080/"
 700         },
 701         "id": "VBdr2epGu3aq",
 702         "outputId": "08bfc16c-0e6b-43d8-8fa3-97b84c124bfb"
 703       },
 704       "source": [
 705         "## tensorflow dataset\n",
 706         "batch_size = 256\n",
 707         "buffer_size = 10000\n",
 708         "\n",
 709         "train_ss = tf.data.Dataset.from_tensor_slices((x_train_ss, y_train_ss))\n",
 710         "train_ss = train_ss.cache().shuffle(buffer_size).batch(batch_size).repeat()\n",
 711         "\n",
 712         "val_ss = tf.data.Dataset.from_tensor_slices((x_val_ss, y_val_ss))\n",
 713         "val_ss = val_ss.cache().shuffle(buffer_size).batch(batch_size).repeat()\n",
 714         "\n",
 715         "print(train_ss)\n",
 716         "print(val_ss)"
 717       ],
 718       "execution_count": null,
 719       "outputs": [
 720         {
 721           "output_type": "stream",
 722           "name": "stdout",
 723           "text": [
 724             "<RepeatDataset shapes: ((None, 120, 7), (None,)), types: (tf.float64, tf.float64)>\n",
 725             "<RepeatDataset shapes: ((None, 120, 7), (None,)), types: (tf.float64, tf.float64)>\n"
 726           ]
 727         }
 728       ]
 729     },
 730     {
 731       "cell_type": "markdown",
 732       "metadata": {
 733         "id": "F-ntkTP2njMy"
 734       },
 735       "source": [
 736         "We used a custom loss function to evaluate the model:"
 737       ]
 738     },
 739     {
 740       "cell_type": "code",
 741       "metadata": {
 742         "id": "9eQpwUyGglu_"
 743       },
 744       "source": [
 745         "def root_mean_squared_error(y_true, y_pred):\n",
 746         "        return K.sqrt(K.mean(K.square(y_pred - y_true))) "
 747       ],
 748       "execution_count": null,
 749       "outputs": []
 750     },
 751     {
 752       "cell_type": "markdown",
 753       "metadata": {
 754         "id": "kTwy7LdqoGpM"
 755       },
 756       "source": [
 757         "We are building a single step LSTM model for training data with dropout 0.3 and we used ADAM optimizers."
 758       ]
 759     },
 760     {
 761       "cell_type": "code",
 762       "metadata": {
 763         "colab": {
 764           "base_uri": "https://localhost:8080/",
 765           "height": 1000
 766         },
 767         "id": "1cKtTAzqyiyL",
 768         "outputId": "f6ec7f52-8cf8-4e81-82c8-8ca59e89f429"
 769       },
 770       "source": [
 771         "from keras.layers import Activation, Dense, Dropout\n",
 772         "from keras.utils.vis_utils import plot_model\n",
 773         "### Modelling using LSTM\n",
 774         "steps = 50\n",
 775         "\n",
 776         "EPOCHS =20\n",
 777         "\n",
 778         "single_step_model = tf.keras.models.Sequential()\n",
 779         "\n",
 780         "single_step_model.add(tf.keras.layers.LSTM(32, return_sequences=False, input_shape = x_train_ss.shape[-2:]))\n",
 781         "single_step_model.add(tf.keras.layers.Dropout(0.3))\n",
 782         "single_step_model.add(tf.keras.layers.Dense(1))\n",
 783         "single_step_model.compile(optimizer = tf.keras.optimizers.Adam(), loss = 'mae',metrics=[tf.keras.metrics.RootMeanSquaredError(name='rmse')])\n",
 784         "#single_step_model.compile(loss='mse', optimizer='rmsprop')\n",
 785         "single_step_model_history = single_step_model.fit(train_ss, epochs = EPOCHS , \n",
 786         "                                                  steps_per_epoch =steps, validation_data = val_ss,\n",
 787         "                                                  validation_steps = 50)\n",
 788         "single_step_model.summary()\n",
 789         "plot_model(single_step_model, to_file='/content/drive/MyDrive/Failure/lstm/LSTM.png', show_shapes=True, show_layer_names=True)\n"
 790       ],
 791       "execution_count": null,
 792       "outputs": [
 793         {
 794           "output_type": "stream",
 795           "name": "stdout",
 796           "text": [
 797             "Epoch 1/20\n",
 798             "50/50 [==============================] - 9s 136ms/step - loss: 0.0899 - rmse: 0.1852 - val_loss: 0.0184 - val_rmse: 0.0281\n",
 799             "Epoch 2/20\n",
 800             "50/50 [==============================] - 6s 126ms/step - loss: 0.0069 - rmse: 0.0119 - val_loss: 0.0130 - val_rmse: 0.0175\n",
 801             "Epoch 3/20\n",
 802             "50/50 [==============================] - 6s 125ms/step - loss: 0.0047 - rmse: 0.0078 - val_loss: 0.0058 - val_rmse: 0.0093\n",
 803             "Epoch 4/20\n",
 804             "50/50 [==============================] - 6s 124ms/step - loss: 0.0034 - rmse: 0.0046 - val_loss: 0.0103 - val_rmse: 0.0105\n",
 805             "Epoch 5/20\n",
 806             "50/50 [==============================] - 6s 127ms/step - loss: 0.0083 - rmse: 0.0106 - val_loss: 0.0089 - val_rmse: 0.0089\n",
 807             "Epoch 6/20\n",
 808             "50/50 [==============================] - 6s 126ms/step - loss: 0.0057 - rmse: 0.0070 - val_loss: 0.0018 - val_rmse: 0.0021\n",
 809             "Epoch 7/20\n",
 810             "50/50 [==============================] - 6s 125ms/step - loss: 0.0039 - rmse: 0.0044 - val_loss: 0.0052 - val_rmse: 0.0054\n",
 811             "Epoch 8/20\n",
 812             "50/50 [==============================] - 6s 124ms/step - loss: 0.0038 - rmse: 0.0041 - val_loss: 0.0036 - val_rmse: 0.0038\n",
 813             "Epoch 9/20\n",
 814             "50/50 [==============================] - 6s 127ms/step - loss: 0.0035 - rmse: 0.0038 - val_loss: 0.0035 - val_rmse: 0.0038\n",
 815             "Epoch 10/20\n",
 816             "50/50 [==============================] - 6s 126ms/step - loss: 0.0033 - rmse: 0.0039 - val_loss: 0.0015 - val_rmse: 0.0021\n",
 817             "Epoch 11/20\n",
 818             "50/50 [==============================] - 6s 125ms/step - loss: 0.0069 - rmse: 0.0082 - val_loss: 0.0030 - val_rmse: 0.0033\n",
 819             "Epoch 12/20\n",
 820             "50/50 [==============================] - 6s 126ms/step - loss: 0.0057 - rmse: 0.0083 - val_loss: 0.0044 - val_rmse: 0.0046\n",
 821             "Epoch 13/20\n",
 822             "50/50 [==============================] - 6s 126ms/step - loss: 0.0061 - rmse: 0.0073 - val_loss: 0.0109 - val_rmse: 0.0110\n",
 823             "Epoch 14/20\n",
 824             "50/50 [==============================] - 6s 125ms/step - loss: 0.0071 - rmse: 0.0088 - val_loss: 9.9332e-04 - val_rmse: 0.0014\n",
 825             "Epoch 15/20\n",
 826             "50/50 [==============================] - 6s 126ms/step - loss: 0.0031 - rmse: 0.0034 - val_loss: 0.0026 - val_rmse: 0.0028\n",
 827             "Epoch 16/20\n",
 828             "50/50 [==============================] - 6s 127ms/step - loss: 0.0030 - rmse: 0.0031 - val_loss: 0.0032 - val_rmse: 0.0034\n",
 829             "Epoch 17/20\n",
 830             "50/50 [==============================] - 6s 126ms/step - loss: 0.0029 - rmse: 0.0031 - val_loss: 0.0026 - val_rmse: 0.0028\n",
 831             "Epoch 18/20\n",
 832             "50/50 [==============================] - 6s 128ms/step - loss: 0.0028 - rmse: 0.0030 - val_loss: 0.0030 - val_rmse: 0.0032\n",
 833             "Epoch 19/20\n",
 834             "50/50 [==============================] - 6s 126ms/step - loss: 0.0028 - rmse: 0.0031 - val_loss: 0.0021 - val_rmse: 0.0024\n",
 835             "Epoch 20/20\n",
 836             "50/50 [==============================] - 6s 127ms/step - loss: 0.0026 - rmse: 0.0029 - val_loss: 0.0027 - val_rmse: 0.0030\n",
 837             "Model: \"sequential\"\n",
 838             "_________________________________________________________________\n",
 839             "Layer (type)                 Output Shape              Param #   \n",
 840             "=================================================================\n",
 841             "lstm (LSTM)                  (None, 32)                5120      \n",
 842             "_________________________________________________________________\n",
 843             "dense (Dense)                (None, 1)                 33        \n",
 844             "=================================================================\n",
 845             "Total params: 5,153\n",
 846             "Trainable params: 5,153\n",
 847             "Non-trainable params: 0\n",
 848             "_________________________________________________________________\n"
 849           ]
 850         },
 851         {
 852           "output_type": "execute_result",
 853           "data": {
 854             "image/png": "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\n",
 855             "text/plain": [
 856               "<IPython.core.display.Image object>"
 857             ]
 858           },
 859           "metadata": {},
 860           "execution_count": 17
 861         }
 862       ]
 863     },
 864     {
 865       "cell_type": "markdown",
 866       "metadata": {
 867         "id": "ch-I96MromT8"
 868       },
 869       "source": [
 870         "We defined the `plot_loss` function to plot the train and test loss"
 871       ]
 872     },
 873     {
 874       "cell_type": "code",
 875       "metadata": {
 876         "colab": {
 877           "base_uri": "https://localhost:8080/",
 878           "height": 281
 879         },
 880         "id": "Pgev0dgzzBVx",
 881         "outputId": "c8794bd3-d704-4244-8f32-7be7d73f0890"
 882       },
 883       "source": [
 884         "## plot train test loss \n",
 885         "\n",
 886         "def plot_loss(history , title):\n",
 887         "  loss = history.history['loss']\n",
 888         "  val_loss = history.history['val_loss']\n",
 889         "\n",
 890         "  epochs = range(len(loss))\n",
 891         "  plt.figure()\n",
 892         "  plt.plot(epochs, loss , 'b' , label = 'Train Loss')\n",
 893         "  plt.plot(epochs, val_loss , 'r' , label = 'Validation Loss')\n",
 894         "  plt.title(title)\n",
 895         "  plt.legend()\n",
 896         "  plt.grid()\n",
 897         "  plt.show()\n",
 898         "\n",
 899         "plot_loss(single_step_model_history , 'Single Step Training and validation loss')"
 900       ],
 901       "execution_count": null,
 902       "outputs": [
 903         {
 904           "output_type": "display_data",
 905           "data": {
 906             "image/png": "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\n",
 907             "text/plain": [
 908               "<Figure size 432x288 with 1 Axes>"
 909             ]
 910           },
 911           "metadata": {
 912             "needs_background": "light"
 913           }
 914         }
 915       ]
 916     },
 917     {
 918       "cell_type": "code",
 919       "metadata": {
 920         "colab": {
 921           "base_uri": "https://localhost:8080/",
 922           "height": 281
 923         },
 924         "id": "EnYf6j4okEoC",
 925         "outputId": "06250b6b-9899-4eb1-af6c-14afd3ae70f8"
 926       },
 927       "source": [
 928         "## plot train test loss \n",
 929         "\n",
 930         "def plot_loss(history , title):\n",
 931         "  loss = history.history['rmse']\n",
 932         "  val_loss = history.history['val_rmse']\n",
 933         "\n",
 934         "  epochs = range(len(loss))\n",
 935         "  plt.figure()\n",
 936         "  plt.plot(epochs, loss , 'b' , label = 'Train RMSE')\n",
 937         "  plt.plot(epochs, val_loss , 'r' , label = 'Validation RMSE')\n",
 938         "  plt.title(title)\n",
 939         "  plt.legend()\n",
 940         "  plt.grid()\n",
 941         "  plt.show()\n",
 942         "\n",
 943         "plot_loss(single_step_model_history , 'Single Step Training and validation loss')"
 944       ],
 945       "execution_count": null,
 946       "outputs": [
 947         {
 948           "output_type": "display_data",
 949           "data": {
 950             "image/png": "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\n",
 951             "text/plain": [
 952               "<Figure size 432x288 with 1 Axes>"
 953             ]
 954           },
 955           "metadata": {
 956             "needs_background": "light"
 957           }
 958         }
 959       ]
 960     },
 961     {
 962       "cell_type": "markdown",
 963       "metadata": {
 964         "id": "tBNraVXSqqfH"
 965       },
 966       "source": [
 967         "We defined a function `create_time_steps` to create time steps and function `plot_time_series` to plot the time series data"
 968       ]
 969     },
 970     {
 971       "cell_type": "code",
 972       "metadata": {
 973         "id": "WMegV8mNAwe_"
 974       },
 975       "source": [
 976         "### fucntion to create time steps\n",
 977         "def create_time_steps(length):\n",
 978         "  return list(range(-length,0))\n",
 979         "\n",
 980         "### function to plot time series data\n",
 981         "\n",
 982         "def plot_time_series(plot_data, delta , title):\n",
 983         "  labels = [\"History\" , 'True Future' , 'Model Predcited']\n",
 984         "  marker = ['.-' , 'rx' , 'go']\n",
 985         "  time_steps = create_time_steps(plot_data[0].shape[0])\n",
 986         "\n",
 987         "  if delta:\n",
 988         "    future = delta\n",
 989         "  else:\n",
 990         "    future = 0\n",
 991         "  plt.title(title)\n",
 992         "  for i , x in enumerate(plot_data):\n",
 993         "    if i :\n",
 994         "      plt.plot(future , plot_data[i] , marker[i], markersize = 10 , label = labels[i])\n",
 995         "    else:\n",
 996         "      plt.plot(time_steps, plot_data[i].flatten(), marker[i], label = labels[i])\n",
 997         "  plt.legend()\n",
 998         "  plt.xlim([time_steps[0], (future+5) *2])\n",
 999         "\n",
1000         "  plt.xlabel('Time_Step')\n",
1001         "  return plt"
1002       ],
1003       "execution_count": null,
1004       "outputs": []
1005     },
1006     {
1007       "cell_type": "code",
1008       "metadata": {
1009         "id": "q99i2c-9XKF3"
1010       },
1011       "source": [
1012         "### Moving window average\n",
1013         "\n",
1014         "def MWA(history):\n",
1015         "  return np.mean(history)"
1016       ],
1017       "execution_count": null,
1018       "outputs": []
1019     },
1020     {
1021       "cell_type": "markdown",
1022       "metadata": {
1023         "id": "kX1mtI44q4U7"
1024       },
1025       "source": [
1026         "We plotted the time series and predicted values"
1027       ]
1028     },
1029     {
1030       "cell_type": "code",
1031       "metadata": {
1032         "colab": {
1033           "base_uri": "https://localhost:8080/",
1034           "height": 1000
1035         },
1036         "id": "xFJT1rZDAUVL",
1037         "outputId": "37fe707b-40d1-476b-c668-631789537c57"
1038       },
1039       "source": [
1040         "# plot time series and predicted values\n",
1041         "\n",
1042         "for x, y in val_ss.take(5):\n",
1043         "  plot = plot_time_series([x[0][:, 1].numpy(), y[0].numpy(),\n",
1044         "                    single_step_model.predict(x)[0]], 12,\n",
1045         "                   'Single Step Prediction')\n",
1046         "  plot.show()"
1047       ],
1048       "execution_count": null,
1049       "outputs": [
1050         {
1051           "output_type": "display_data",
1052           "data": {
1053             "image/png": "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\n",
1054             "text/plain": [
1055               "<Figure size 432x288 with 1 Axes>"
1056             ]
1057           },
1058           "metadata": {
1059             "needs_background": "light"
1060           }
1061         },
1062         {
1063           "output_type": "display_data",
1064           "data": {
1065             "image/png": "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\n",
1066             "text/plain": [
1067               "<Figure size 432x288 with 1 Axes>"
1068             ]
1069           },
1070           "metadata": {
1071             "needs_background": "light"
1072           }
1073         },
1074         {
1075           "output_type": "display_data",
1076           "data": {
1077             "image/png": "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\n",
1078             "text/plain": [
1079               "<Figure size 432x288 with 1 Axes>"
1080             ]
1081           },
1082           "metadata": {
1083             "needs_background": "light"
1084           }
1085         },
1086         {
1087           "output_type": "display_data",
1088           "data": {
1089             "image/png": "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\n",
1090             "text/plain": [
1091               "<Figure size 432x288 with 1 Axes>"
1092             ]
1093           },
1094           "metadata": {
1095             "needs_background": "light"
1096           }
1097         },
1098         {
1099           "output_type": "display_data",
1100           "data": {
1101             "image/png": "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\n",
1102             "text/plain": [
1103               "<Figure size 432x288 with 1 Axes>"
1104             ]
1105           },
1106           "metadata": {
1107             "needs_background": "light"
1108           }
1109         }
1110       ]
1111     },
1112     {
1113       "cell_type": "markdown",
1114       "metadata": {
1115         "id": "_KXWQVmyCSix"
1116       },
1117       "source": [
1118         "# **MultiStep Forcasting**"
1119       ]
1120     },
1121     {
1122       "cell_type": "markdown",
1123       "metadata": {
1124         "id": "5ghCNUvcrBMi"
1125       },
1126       "source": [
1127         "We splitted the data in the form of training and validation for multistep forcasting:"
1128       ]
1129     },
1130     {
1131       "cell_type": "code",
1132       "metadata": {
1133         "colab": {
1134           "base_uri": "https://localhost:8080/"
1135         },
1136         "id": "Lu7m2Rr4AbMK",
1137         "outputId": "4aeabbf6-4ca3-4340-d1b1-e94ad1f78faf"
1138       },
1139       "source": [
1140         "future_target = 72 # 72 future values\n",
1141         "x_train_multi, y_train_multi = mutlivariate_data(features, features[:, 1], 0,\n",
1142         "                                                 train_split, history,\n",
1143         "                                                 future_target, STEP)\n",
1144         "x_val_multi, y_val_multi = mutlivariate_data(features, features[:, 1],\n",
1145         "                                             train_split, None, history,\n",
1146         "                                             future_target, STEP)\n",
1147         "\n",
1148         "print(x_train_multi.shape)\n",
1149         "print(y_train_multi.shape)"
1150       ],
1151       "execution_count": null,
1152       "outputs": [
1153         {
1154           "output_type": "stream",
1155           "name": "stdout",
1156           "text": [
1157             "(140880, 120, 7)\n",
1158             "(140880, 72)\n"
1159           ]
1160         }
1161       ]
1162     },
1163     {
1164       "cell_type": "markdown",
1165       "metadata": {
1166         "id": "Rv5mugNBrVdY"
1167       },
1168       "source": [
1169         "The `tf.data.Dataset` API supports writing descriptive and efficient input pipelines. Dataset usage following a common pattern:\n",
1170         "- Creating a source dataset from our input data.\n",
1171         "- Applied dataset transformations to preprocess the data.\n",
1172         "- Iterate over the dataset and process the elements.\n",
1173         "Note: Iteration happens in a streaming fashion, so the full dataset does not need to fit into memory.\n",
1174         "Once we have a dataset, we can apply transformations to prepare the data for our model:"
1175       ]
1176     },
1177     {
1178       "cell_type": "code",
1179       "metadata": {
1180         "id": "GLRv5D16CrHj"
1181       },
1182       "source": [
1183         "#  TF DATASET\n",
1184         "\n",
1185         "train_data_multi = tf.data.Dataset.from_tensor_slices((x_train_multi, y_train_multi))\n",
1186         "train_data_multi = train_data_multi.cache().shuffle(buffer_size).batch(batch_size).repeat()\n",
1187         "\n",
1188         "val_data_multi = tf.data.Dataset.from_tensor_slices((x_val_multi, y_val_multi))\n",
1189         "val_data_multi = val_data_multi.batch(batch_size).repeat()"
1190       ],
1191       "execution_count": null,
1192       "outputs": []
1193     },
1194     {
1195       "cell_type": "code",
1196       "metadata": {
1197         "colab": {
1198           "base_uri": "https://localhost:8080/"
1199         },
1200         "id": "fjXexah9C8yg",
1201         "outputId": "13cf8949-9ebc-4ea9-b5ac-f33c6e146a00"
1202       },
1203       "source": [
1204         "print(train_data_multi)\n",
1205         "print(val_data_multi)"
1206       ],
1207       "execution_count": null,
1208       "outputs": [
1209         {
1210           "output_type": "stream",
1211           "name": "stdout",
1212           "text": [
1213             "<RepeatDataset shapes: ((None, 120, 7), (None, 72)), types: (tf.float64, tf.float64)>\n",
1214             "<RepeatDataset shapes: ((None, 120, 7), (None, 72)), types: (tf.float64, tf.float64)>\n"
1215           ]
1216         }
1217       ]
1218     },
1219     {
1220       "cell_type": "markdown",
1221       "metadata": {
1222         "id": "gMIQq83Ercb2"
1223       },
1224       "source": [
1225         "We created a `multi_step_plot` function to plot between `history` and `true_future` data"
1226       ]
1227     },
1228     {
1229       "cell_type": "code",
1230       "metadata": {
1231         "colab": {
1232           "base_uri": "https://localhost:8080/",
1233           "height": 385
1234         },
1235         "id": "7mtLZ6S-DPU-",
1236         "outputId": "eadda952-c8bc-4b54-e55f-0500d73115c1"
1237       },
1238       "source": [
1239         "#plotting function\n",
1240         "def multi_step_plot(history, true_future, prediction):\n",
1241         "  plt.figure(figsize=(12, 6))\n",
1242         "  num_in = create_time_steps(len(history))\n",
1243         "  num_out = len(true_future)\n",
1244         "  plt.grid()\n",
1245         "  plt.plot(num_in, np.array(history[:, 1]), label='History')\n",
1246         "  plt.plot(np.arange(num_out)/STEP, np.array(true_future), 'bo',\n",
1247         "           label='True Future')\n",
1248         "  if prediction.any():\n",
1249         "    plt.plot(np.arange(num_out)/STEP, np.array(prediction), 'ro',\n",
1250         "             label='Predicted Future')\n",
1251         "  plt.legend(loc='upper left')\n",
1252         "  plt.show()\n",
1253         "  \n",
1254         "\n",
1255         "\n",
1256         "for x, y in train_data_multi.take(1):\n",
1257         "  multi_step_plot(x[0], y[0], np.array([0]))"
1258       ],
1259       "execution_count": null,
1260       "outputs": [
1261         {
1262           "output_type": "display_data",
1263           "data": {
1264             "image/png": "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\n",
1265             "text/plain": [
1266               "<Figure size 864x432 with 1 Axes>"
1267             ]
1268           },
1269           "metadata": {
1270             "needs_background": "light"
1271           }
1272         }
1273       ]
1274     },
1275     {
1276       "cell_type": "markdown",
1277       "metadata": {
1278         "id": "pJ7kryvdrqcg"
1279       },
1280       "source": [
1281         "We are building a single step LSTM model for training data with dropout 0.3 and we used ADAM optimizers."
1282       ]
1283     },
1284     {
1285       "cell_type": "code",
1286       "metadata": {
1287         "colab": {
1288           "base_uri": "https://localhost:8080/"
1289         },
1290         "id": "snN_Flr5DWQN",
1291         "outputId": "449c5036-923b-4a60-8e6d-dad286ef1c59"
1292       },
1293       "source": [
1294         "multi_step_model = tf.keras.models.Sequential()\n",
1295         "multi_step_model.add(tf.keras.layers.LSTM(32,\n",
1296         "                                          return_sequences=True,\n",
1297         "                                          input_shape=x_train_multi.shape[-2:]))\n",
1298         "multi_step_model.add(tf.keras.layers.LSTM(16, activation='relu'))\n",
1299         "#aDD dropout layer (0.3)\n",
1300         "multi_step_model.add(tf.keras.layers.Dense(72)) # for 72 outputs\n",
1301         "\n",
1302         "multi_step_model.compile(optimizer=tf.keras.optimizers.RMSprop(clipvalue=1.0), loss='mae',metrics=[tf.keras.metrics.RootMeanSquaredError(name='rmse')])\n",
1303         "\n",
1304         "multi_step_history = multi_step_model.fit(train_data_multi, epochs=EPOCHS,\n",
1305         "                                          steps_per_epoch=steps,\n",
1306         "                                          validation_data=val_data_multi,\n",
1307         "                                          validation_steps=50)"
1308       ],
1309       "execution_count": null,
1310       "outputs": [
1311         {
1312           "output_type": "stream",
1313           "name": "stdout",
1314           "text": [
1315             "Epoch 1/20\n",
1316             "50/50 [==============================] - 14s 226ms/step - loss: 0.2044 - rmse: 0.2723 - val_loss: 0.0922 - val_rmse: 0.1511\n",
1317             "Epoch 2/20\n",
1318             "50/50 [==============================] - 11s 215ms/step - loss: 0.0574 - rmse: 0.1058 - val_loss: 0.0394 - val_rmse: 0.0519\n",
1319             "Epoch 3/20\n",
1320             "50/50 [==============================] - 11s 216ms/step - loss: 0.0240 - rmse: 0.0351 - val_loss: 0.0156 - val_rmse: 0.0187\n",
1321             "Epoch 4/20\n",
1322             "50/50 [==============================] - 11s 214ms/step - loss: 0.0182 - rmse: 0.0226 - val_loss: 0.0054 - val_rmse: 0.0068\n",
1323             "Epoch 5/20\n",
1324             "50/50 [==============================] - 11s 212ms/step - loss: 0.0164 - rmse: 0.0199 - val_loss: 0.0096 - val_rmse: 0.0110\n",
1325             "Epoch 6/20\n",
1326             "50/50 [==============================] - 11s 219ms/step - loss: 0.0148 - rmse: 0.0179 - val_loss: 0.0129 - val_rmse: 0.0157\n",
1327             "Epoch 7/20\n",
1328             "50/50 [==============================] - 11s 217ms/step - loss: 0.0137 - rmse: 0.0163 - val_loss: 0.0071 - val_rmse: 0.0084\n",
1329             "Epoch 8/20\n",
1330             "50/50 [==============================] - 11s 216ms/step - loss: 0.0138 - rmse: 0.0162 - val_loss: 0.0090 - val_rmse: 0.0100\n",
1331             "Epoch 9/20\n",
1332             "50/50 [==============================] - 11s 216ms/step - loss: 0.0131 - rmse: 0.0152 - val_loss: 0.0081 - val_rmse: 0.0092\n",
1333             "Epoch 10/20\n",
1334             "50/50 [==============================] - 11s 218ms/step - loss: 0.0126 - rmse: 0.0144 - val_loss: 0.0081 - val_rmse: 0.0090\n",
1335             "Epoch 11/20\n",
1336             "50/50 [==============================] - 11s 214ms/step - loss: 0.0121 - rmse: 0.0137 - val_loss: 0.0082 - val_rmse: 0.0090\n",
1337             "Epoch 12/20\n",
1338             "50/50 [==============================] - 11s 215ms/step - loss: 0.0191 - rmse: 0.0297 - val_loss: 0.0055 - val_rmse: 0.0068\n",
1339             "Epoch 13/20\n",
1340             "50/50 [==============================] - 11s 218ms/step - loss: 0.0112 - rmse: 0.0133 - val_loss: 0.0067 - val_rmse: 0.0080\n",
1341             "Epoch 14/20\n",
1342             "50/50 [==============================] - 11s 214ms/step - loss: 0.0104 - rmse: 0.0124 - val_loss: 0.0039 - val_rmse: 0.0048\n",
1343             "Epoch 15/20\n",
1344             "50/50 [==============================] - 11s 216ms/step - loss: 0.0105 - rmse: 0.0127 - val_loss: 0.0049 - val_rmse: 0.0058\n",
1345             "Epoch 16/20\n",
1346             "50/50 [==============================] - 11s 215ms/step - loss: 0.0102 - rmse: 0.0122 - val_loss: 0.0029 - val_rmse: 0.0035\n",
1347             "Epoch 17/20\n",
1348             "50/50 [==============================] - 11s 216ms/step - loss: 0.0098 - rmse: 0.0117 - val_loss: 0.0043 - val_rmse: 0.0052\n",
1349             "Epoch 18/20\n",
1350             "50/50 [==============================] - 11s 217ms/step - loss: 0.0088 - rmse: 0.0108 - val_loss: 0.0076 - val_rmse: 0.0094\n",
1351             "Epoch 19/20\n",
1352             "50/50 [==============================] - 11s 219ms/step - loss: 0.0090 - rmse: 0.0108 - val_loss: 0.0141 - val_rmse: 0.0143\n",
1353             "Epoch 20/20\n",
1354             "50/50 [==============================] - 11s 215ms/step - loss: 0.0092 - rmse: 0.0109 - val_loss: 0.0135 - val_rmse: 0.0139\n"
1355           ]
1356         }
1357       ]
1358     },
1359     {
1360       "cell_type": "code",
1361       "metadata": {
1362         "colab": {
1363           "base_uri": "https://localhost:8080/",
1364           "height": 281
1365         },
1366         "id": "Ay5m27doDsTt",
1367         "outputId": "803927cb-1be0-4f8b-e31e-2dcb68a7e217"
1368       },
1369       "source": [
1370         "plot_loss(multi_step_history, 'Multi-Step Training and validation loss')"
1371       ],
1372       "execution_count": null,
1373       "outputs": [
1374         {
1375           "output_type": "display_data",
1376           "data": {
1377             "image/png": "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\n",
1378             "text/plain": [
1379               "<Figure size 432x288 with 1 Axes>"
1380             ]
1381           },
1382           "metadata": {
1383             "needs_background": "light"
1384           }
1385         }
1386       ]
1387     },
1388     {
1389       "cell_type": "code",
1390       "metadata": {
1391         "colab": {
1392           "base_uri": "https://localhost:8080/",
1393           "height": 1000
1394         },
1395         "id": "6ZFP49W4D2wp",
1396         "outputId": "b810ee24-c241-45ce-e60c-c002d81270d0"
1397       },
1398       "source": [
1399         "for x, y in val_data_multi.take(5):\n",
1400         "  multi_step_plot(x[0], y[0], multi_step_model.predict(x)[0])"
1401       ],
1402       "execution_count": null,
1403       "outputs": [
1404         {
1405           "output_type": "display_data",
1406           "data": {
1407             "image/png": "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\n",
1408             "text/plain": [
1409               "<Figure size 864x432 with 1 Axes>"
1410             ]
1411           },
1412           "metadata": {
1413             "needs_background": "light"
1414           }
1415         },
1416         {
1417           "output_type": "display_data",
1418           "data": {
1419             "image/png": "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\n",
1420             "text/plain": [
1421               "<Figure size 864x432 with 1 Axes>"
1422             ]
1423           },
1424           "metadata": {
1425             "needs_background": "light"
1426           }
1427         },
1428         {
1429           "output_type": "display_data",
1430           "data": {
1431             "image/png": "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\n",
1432             "text/plain": [
1433               "<Figure size 864x432 with 1 Axes>"
1434             ]
1435           },
1436           "metadata": {
1437             "needs_background": "light"
1438           }
1439         },
1440         {
1441           "output_type": "display_data",
1442           "data": {
1443             "image/png": "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\n",
1444             "text/plain": [
1445               "<Figure size 864x432 with 1 Axes>"
1446             ]
1447           },
1448           "metadata": {
1449             "needs_background": "light"
1450           }
1451         },
1452         {
1453           "output_type": "display_data",
1454           "data": {
1455             "image/png": "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\n",
1456             "text/plain": [
1457               "<Figure size 864x432 with 1 Axes>"
1458             ]
1459           },
1460           "metadata": {
1461             "needs_background": "light"
1462           }
1463         }
1464       ]
1465     },
1466     {
1467       "cell_type": "code",
1468       "metadata": {
1469         "id": "DNKMjVoAVqZP",
1470         "colab": {
1471           "base_uri": "https://localhost:8080/"
1472         },
1473         "outputId": "8639928a-99de-4dbd-9394-34b651da23ac"
1474       },
1475       "source": [
1476         "scores = multi_step_model.evaluate(x_train_multi, y_train_multi, verbose=1, batch_size=200)\n",
1477         "print('MAE: {}'.format(scores[1]))"
1478       ],
1479       "execution_count": null,
1480       "outputs": [
1481         {
1482           "output_type": "stream",
1483           "name": "stdout",
1484           "text": [
1485             "705/705 [==============================] - 36s 51ms/step - loss: 0.0147 - rmse: 0.0161\n",
1486             "MAE: 0.016090337187051773\n"
1487           ]
1488         }
1489       ]
1490     },
1491     {
1492       "cell_type": "code",
1493       "metadata": {
1494         "colab": {
1495           "base_uri": "https://localhost:8080/"
1496         },
1497         "id": "oLqijoE9W-Gm",
1498         "outputId": "ccb080da-2676-4cf8-d3a3-968bc5cb6c94"
1499       },
1500       "source": [
1501         "scores_test = multi_step_model.evaluate(x_val_multi, y_val_multi, verbose=1, batch_size=200)\n",
1502         "print('MAE: {}'.format(scores[1]))"
1503       ],
1504       "execution_count": null,
1505       "outputs": [
1506         {
1507           "output_type": "stream",
1508           "name": "stdout",
1509           "text": [
1510             "174/174 [==============================] - 9s 51ms/step - loss: 0.0142 - rmse: 0.0151\n",
1511             "MAE: 0.016090337187051773\n"
1512           ]
1513         }
1514       ]
1515     }
1516   ]
1517 }