Air_Quality_Index.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "Air_Quality_Index.ipynb",
      "provenance": [],
      "authorship_tag": "ABX9TyOzsK8B+9wdHUeq3fo0n2fP",
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/pravali96/Artificial-Neural-Networks/blob/main/Air_Quality_Index.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ZHGlWwg8OcCK"
      },
      "source": [
        "import pandas as pd\n",
        "import numpy as np\n",
        "import tensorflow as tf\n",
        "from tensorflow import keras\n",
        "!pip install -q -U keras-tuner\n",
        "import keras_tuner as kt"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "resources": {
            "http://localhost:8080/nbextensions/google.colab/files.js": {
              "data": "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",
              "ok": true,
              "headers": [
                [
                  "content-type",
                  "application/javascript"
                ]
              ],
              "status": 200,
              "status_text": ""
            }
          },
          "base_uri": "https://localhost:8080/",
          "height": 73
        },
        "id": "P4IJeVgqVqL3",
        "outputId": "4120f172-ccf9-4068-dc7c-5210d2ca636d"
      },
      "source": [
        "from google.colab import files\n",
        "uploaded=files.upload()\n",
        "import io"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/html": [
              "\n",
              "     <input type=\"file\" id=\"files-cb5ab353-acbf-4eb2-9c1f-04c422674b6b\" name=\"files[]\" multiple disabled\n",
              "        style=\"border:none\" />\n",
              "     <output id=\"result-cb5ab353-acbf-4eb2-9c1f-04c422674b6b\">\n",
              "      Upload widget is only available when the cell has been executed in the\n",
              "      current browser session. Please rerun this cell to enable.\n",
              "      </output>\n",
              "      <script src=\"/nbextensions/google.colab/files.js\"></script> "
            ],
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "stream",
          "text": [
            "Saving air_quality.csv to air_quality (2).csv\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "dyV01-56V25a"
      },
      "source": [
        "df=pd.read_csv('air_quality.csv')"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 196
        },
        "id": "wYRI8D52V-DS",
        "outputId": "92ab82e8-0d8f-40b8-c391-4d436a6f2b0c"
      },
      "source": [
        "df.head()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "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>T</th>\n",
              "      <th>TM</th>\n",
              "      <th>Tm</th>\n",
              "      <th>SLP</th>\n",
              "      <th>H</th>\n",
              "      <th>VV</th>\n",
              "      <th>V</th>\n",
              "      <th>VM</th>\n",
              "      <th>PM 2.5</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>7.4</td>\n",
              "      <td>9.8</td>\n",
              "      <td>4.8</td>\n",
              "      <td>1017.6</td>\n",
              "      <td>93.0</td>\n",
              "      <td>0.5</td>\n",
              "      <td>4.3</td>\n",
              "      <td>9.4</td>\n",
              "      <td>219.720833</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>7.8</td>\n",
              "      <td>12.7</td>\n",
              "      <td>4.4</td>\n",
              "      <td>1018.5</td>\n",
              "      <td>87.0</td>\n",
              "      <td>0.6</td>\n",
              "      <td>4.4</td>\n",
              "      <td>11.1</td>\n",
              "      <td>182.187500</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>6.7</td>\n",
              "      <td>13.4</td>\n",
              "      <td>2.4</td>\n",
              "      <td>1019.4</td>\n",
              "      <td>82.0</td>\n",
              "      <td>0.6</td>\n",
              "      <td>4.8</td>\n",
              "      <td>11.1</td>\n",
              "      <td>154.037500</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>8.6</td>\n",
              "      <td>15.5</td>\n",
              "      <td>3.3</td>\n",
              "      <td>1018.7</td>\n",
              "      <td>72.0</td>\n",
              "      <td>0.8</td>\n",
              "      <td>8.1</td>\n",
              "      <td>20.6</td>\n",
              "      <td>223.208333</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>12.4</td>\n",
              "      <td>20.9</td>\n",
              "      <td>4.4</td>\n",
              "      <td>1017.3</td>\n",
              "      <td>61.0</td>\n",
              "      <td>1.3</td>\n",
              "      <td>8.7</td>\n",
              "      <td>22.2</td>\n",
              "      <td>200.645833</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "      T    TM   Tm     SLP     H   VV    V    VM      PM 2.5\n",
              "0   7.4   9.8  4.8  1017.6  93.0  0.5  4.3   9.4  219.720833\n",
              "1   7.8  12.7  4.4  1018.5  87.0  0.6  4.4  11.1  182.187500\n",
              "2   6.7  13.4  2.4  1019.4  82.0  0.6  4.8  11.1  154.037500\n",
              "3   8.6  15.5  3.3  1018.7  72.0  0.8  8.1  20.6  223.208333\n",
              "4  12.4  20.9  4.4  1017.3  61.0  1.3  8.7  22.2  200.645833"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 4
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "tFR8c1KOWCG5",
        "outputId": "d7af7296-1ab4-4c1a-8def-ee3425484843"
      },
      "source": [
        "df.isnull().sum()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "T         0\n",
              "TM        0\n",
              "Tm        0\n",
              "SLP       0\n",
              "H         0\n",
              "VV        0\n",
              "V         0\n",
              "VM        0\n",
              "PM 2.5    1\n",
              "dtype: int64"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 5
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "lsxi3dQXWD_J"
      },
      "source": [
        "df=df.dropna()"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "1OG8qcEpWFc-"
      },
      "source": [
        "y=df.iloc[:,-1]\n",
        "X=df.iloc[:,:-1]"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "DLM4Y8N0W3TA"
      },
      "source": [
        "# Creating Train test split\n",
        "from sklearn.model_selection import train_test_split\n",
        "X_train, X_test, y_train, y_test = train_test_split(X,y, test_size=0.2, random_state=0)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ARzb59a6WhKq"
      },
      "source": [
        "from sklearn.preprocessing import StandardScaler\n",
        "sc=StandardScaler()\n",
        "X_train=sc.fit_transform(X_train)\n",
        "X_test=sc.transform(X_test)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "D5_76_TzWm_P"
      },
      "source": [
        "def model_builder(hp):\n",
        "  model = keras.Sequential()\n",
        "  for i in range(hp.Int('num_layers',2,20)):\n",
        "    model.add(keras.layers.Dense(units=hp.Int('units_' + str(i),\n",
        "                                            min_value=32,\n",
        "                                            max_value=512,\n",
        "                                            step=32),\n",
        "                               activation='relu'))\n",
        "  # Tune the number of units in the first Dense layer\n",
        "  # Choose an optimal value between 32-512\n",
        "  model.add(keras.layers.Dense(1, activation='linear'))\n",
        "  # Tune the learning rate for the optimizer\n",
        "  # Choose an optimal value from 0.01, 0.001, or 0.0001\n",
        "  hp_learning_rate = hp.Choice('learning_rate', values=[1e-2, 1e-3, 1e-4])\n",
        "  model.compile(optimizer=keras.optimizers.Adam(learning_rate=hp_learning_rate),\n",
        "                loss='mean_absolute_error',\n",
        "                metrics=['mean_absolute_error'])\n",
        "\n",
        "  return model"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "jp2bQsHMdSfA",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "0e4913fb-6994-4eb5-a4d3-960455dfe39d"
      },
      "source": [
        "from keras_tuner import RandomSearch\n",
        "tuner = RandomSearch(\n",
        "    model_builder,\n",
        "    objective='val_mean_absolute_error',\n",
        "    max_trials=5,\n",
        "    executions_per_trial=3,\n",
        "    directory='project',\n",
        "    project_name='Air Quality Index')"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "INFO:tensorflow:Reloading Oracle from existing project project/Air Quality Index/oracle.json\n",
            "INFO:tensorflow:Reloading Tuner from project/Air Quality Index/tuner0.json\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "wofEREcBZWdk",
        "outputId": "3df5ec84-e13a-4214-f913-cc1885f5256c"
      },
      "source": [
        "tuner.search_space_summary"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<bound method BaseTuner.search_space_summary of <keras_tuner.tuners.randomsearch.RandomSearch object at 0x7f1efc401150>>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 12
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "aPm4N15iZtMw",
        "outputId": "65916d23-d3ec-48c1-f183-0f4bdde89b1e"
      },
      "source": [
        "tuner.search(X_train, y_train,\n",
        "             epochs=5,\n",
        "             validation_data=(X_test, y_test))"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "INFO:tensorflow:Oracle triggered exit\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "RH8Zx0lKZ8XM",
        "outputId": "6190f023-8f6d-44ee-f6a1-761211eec30e"
      },
      "source": [
        "tuner.results_summary()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Results summary\n",
            "Results in project/Air Quality Index\n",
            "Showing 10 best trials\n",
            "Objective(name='val_mean_absolute_error', direction='min')\n",
            "Trial summary\n",
            "Hyperparameters:\n",
            "num_layers: 17\n",
            "units_0: 128\n",
            "units_1: 416\n",
            "learning_rate: 0.001\n",
            "units_2: 224\n",
            "units_3: 448\n",
            "units_4: 320\n",
            "units_5: 416\n",
            "units_6: 512\n",
            "units_7: 320\n",
            "units_8: 64\n",
            "units_9: 320\n",
            "units_10: 256\n",
            "units_11: 160\n",
            "units_12: 448\n",
            "units_13: 320\n",
            "units_14: 512\n",
            "units_15: 64\n",
            "units_16: 32\n",
            "Score: 35.00788243611654\n",
            "Trial summary\n",
            "Hyperparameters:\n",
            "num_layers: 4\n",
            "units_0: 416\n",
            "units_1: 416\n",
            "learning_rate: 0.001\n",
            "units_2: 32\n",
            "units_3: 32\n",
            "Score: 41.79108428955078\n",
            "Trial summary\n",
            "Hyperparameters:\n",
            "num_layers: 9\n",
            "units_0: 448\n",
            "units_1: 320\n",
            "learning_rate: 0.0001\n",
            "units_2: 64\n",
            "units_3: 512\n",
            "units_4: 480\n",
            "units_5: 32\n",
            "units_6: 224\n",
            "units_7: 128\n",
            "units_8: 224\n",
            "units_9: 224\n",
            "units_10: 352\n",
            "units_11: 64\n",
            "units_12: 416\n",
            "units_13: 32\n",
            "units_14: 96\n",
            "units_15: 352\n",
            "units_16: 448\n",
            "Score: 46.44004821777344\n",
            "Trial summary\n",
            "Hyperparameters:\n",
            "num_layers: 16\n",
            "units_0: 192\n",
            "units_1: 352\n",
            "learning_rate: 0.0001\n",
            "units_2: 384\n",
            "units_3: 224\n",
            "units_4: 384\n",
            "units_5: 32\n",
            "units_6: 32\n",
            "units_7: 32\n",
            "units_8: 32\n",
            "units_9: 32\n",
            "units_10: 32\n",
            "units_11: 32\n",
            "units_12: 32\n",
            "units_13: 32\n",
            "units_14: 32\n",
            "units_15: 32\n",
            "Score: 50.095480600992836\n",
            "Trial summary\n",
            "Hyperparameters:\n",
            "num_layers: 5\n",
            "units_0: 384\n",
            "units_1: 288\n",
            "learning_rate: 0.0001\n",
            "units_2: 128\n",
            "units_3: 448\n",
            "units_4: 32\n",
            "Score: 51.78632354736328\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ZhLh_HdKaKwb",
        "outputId": "cc48ed74-5c85-4ffd-b1dc-2bd6eb65f9a5"
      },
      "source": [
        "best_model = tuner.get_best_models(num_models=1)[0]\n",
        "loss, mae = best_model.evaluate(X_test, y_test)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "7/7 [==============================] - 1s 8ms/step - loss: 34.4206 - mean_absolute_error: 34.4206\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "aCaXyu3Deazk",
        "outputId": "b278c17b-730b-4091-9295-46b347d445cf"
      },
      "source": [
        "best_model.summary()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Model: \"sequential\"\n",
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "dense (Dense)                (None, 128)               1152      \n",
            "_________________________________________________________________\n",
            "dense_1 (Dense)              (None, 416)               53664     \n",
            "_________________________________________________________________\n",
            "dense_2 (Dense)              (None, 224)               93408     \n",
            "_________________________________________________________________\n",
            "dense_3 (Dense)              (None, 448)               100800    \n",
            "_________________________________________________________________\n",
            "dense_4 (Dense)              (None, 320)               143680    \n",
            "_________________________________________________________________\n",
            "dense_5 (Dense)              (None, 416)               133536    \n",
            "_________________________________________________________________\n",
            "dense_6 (Dense)              (None, 512)               213504    \n",
            "_________________________________________________________________\n",
            "dense_7 (Dense)              (None, 320)               164160    \n",
            "_________________________________________________________________\n",
            "dense_8 (Dense)              (None, 64)                20544     \n",
            "_________________________________________________________________\n",
            "dense_9 (Dense)              (None, 320)               20800     \n",
            "_________________________________________________________________\n",
            "dense_10 (Dense)             (None, 256)               82176     \n",
            "_________________________________________________________________\n",
            "dense_11 (Dense)             (None, 160)               41120     \n",
            "_________________________________________________________________\n",
            "dense_12 (Dense)             (None, 448)               72128     \n",
            "_________________________________________________________________\n",
            "dense_13 (Dense)             (None, 320)               143680    \n",
            "_________________________________________________________________\n",
            "dense_14 (Dense)             (None, 512)               164352    \n",
            "_________________________________________________________________\n",
            "dense_15 (Dense)             (None, 64)                32832     \n",
            "_________________________________________________________________\n",
            "dense_16 (Dense)             (None, 32)                2080      \n",
            "_________________________________________________________________\n",
            "dense_17 (Dense)             (None, 1)                 33        \n",
            "=================================================================\n",
            "Total params: 1,483,649\n",
            "Trainable params: 1,483,649\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "67HMtizwgWSU",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "e9e4423b-cf4c-48c8-dc13-3dcf7bfb7b6c"
      },
      "source": [
        "model_hist=best_model.fit(X_train, y_train, validation_split=0.2, epochs=200, initial_epoch=5)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Epoch 6/200\n",
            "22/22 [==============================] - 2s 26ms/step - loss: 35.3599 - mean_absolute_error: 35.3599 - val_loss: 40.4892 - val_mean_absolute_error: 40.4892\n",
            "Epoch 7/200\n",
            "22/22 [==============================] - 1s 26ms/step - loss: 33.3601 - mean_absolute_error: 33.3601 - val_loss: 37.8190 - val_mean_absolute_error: 37.8190\n",
            "Epoch 8/200\n",
            "22/22 [==============================] - 1s 24ms/step - loss: 34.9237 - mean_absolute_error: 34.9237 - val_loss: 39.6538 - val_mean_absolute_error: 39.6538\n",
            "Epoch 9/200\n",
            "22/22 [==============================] - 1s 24ms/step - loss: 37.3440 - mean_absolute_error: 37.3440 - val_loss: 48.5649 - val_mean_absolute_error: 48.5649\n",
            "Epoch 10/200\n",
            "22/22 [==============================] - 1s 24ms/step - loss: 35.1027 - mean_absolute_error: 35.1027 - val_loss: 37.9327 - val_mean_absolute_error: 37.9327\n",
            "Epoch 11/200\n",
            "22/22 [==============================] - 1s 25ms/step - loss: 34.3867 - mean_absolute_error: 34.3867 - val_loss: 48.2592 - val_mean_absolute_error: 48.2592\n",
            "Epoch 12/200\n",
            "22/22 [==============================] - 1s 25ms/step - loss: 33.6295 - mean_absolute_error: 33.6295 - val_loss: 42.6929 - val_mean_absolute_error: 42.6929\n",
            "Epoch 13/200\n",
            "22/22 [==============================] - 1s 26ms/step - loss: 31.9885 - mean_absolute_error: 31.9885 - val_loss: 36.5988 - val_mean_absolute_error: 36.5988\n",
            "Epoch 14/200\n",
            "22/22 [==============================] - 1s 25ms/step - loss: 33.5614 - mean_absolute_error: 33.5614 - val_loss: 42.5400 - val_mean_absolute_error: 42.5400\n",
            "Epoch 15/200\n",
            "22/22 [==============================] - 1s 25ms/step - loss: 33.7358 - mean_absolute_error: 33.7358 - val_loss: 37.5967 - val_mean_absolute_error: 37.5967\n",
            "Epoch 16/200\n",
            "22/22 [==============================] - 1s 26ms/step - loss: 31.0550 - mean_absolute_error: 31.0550 - val_loss: 38.0223 - val_mean_absolute_error: 38.0223\n",
            "Epoch 17/200\n",
            "22/22 [==============================] - 1s 24ms/step - loss: 29.9084 - mean_absolute_error: 29.9084 - val_loss: 40.3675 - val_mean_absolute_error: 40.3675\n",
            "Epoch 18/200\n",
            "22/22 [==============================] - 1s 24ms/step - loss: 33.5413 - mean_absolute_error: 33.5413 - val_loss: 35.7434 - val_mean_absolute_error: 35.7434\n",
            "Epoch 19/200\n",
            "22/22 [==============================] - 1s 23ms/step - loss: 29.5201 - mean_absolute_error: 29.5201 - val_loss: 36.9964 - val_mean_absolute_error: 36.9964\n",
            "Epoch 20/200\n",
            "22/22 [==============================] - 1s 26ms/step - loss: 28.1849 - mean_absolute_error: 28.1849 - val_loss: 39.6863 - val_mean_absolute_error: 39.6863\n",
            "Epoch 21/200\n",
            "22/22 [==============================] - 1s 25ms/step - loss: 29.8813 - mean_absolute_error: 29.8813 - val_loss: 35.2831 - val_mean_absolute_error: 35.2831\n",
            "Epoch 22/200\n",
            "22/22 [==============================] - 1s 25ms/step - loss: 27.4651 - mean_absolute_error: 27.4651 - val_loss: 38.0576 - val_mean_absolute_error: 38.0576\n",
            "Epoch 23/200\n",
            "22/22 [==============================] - 1s 25ms/step - loss: 26.9022 - mean_absolute_error: 26.9022 - val_loss: 35.2023 - val_mean_absolute_error: 35.2023\n",
            "Epoch 24/200\n",
            "22/22 [==============================] - 1s 26ms/step - loss: 27.7912 - mean_absolute_error: 27.7912 - val_loss: 35.8100 - val_mean_absolute_error: 35.8100\n",
            "Epoch 25/200\n",
            "22/22 [==============================] - 1s 24ms/step - loss: 27.6140 - mean_absolute_error: 27.6140 - val_loss: 34.4791 - val_mean_absolute_error: 34.4791\n",
            "Epoch 26/200\n",
            "22/22 [==============================] - 1s 24ms/step - loss: 32.0584 - mean_absolute_error: 32.0584 - val_loss: 39.0396 - val_mean_absolute_error: 39.0396\n",
            "Epoch 27/200\n",
            "22/22 [==============================] - 1s 25ms/step - loss: 26.5716 - mean_absolute_error: 26.5716 - val_loss: 36.7301 - val_mean_absolute_error: 36.7301\n",
            "Epoch 28/200\n",
            "22/22 [==============================] - 1s 25ms/step - loss: 25.2595 - mean_absolute_error: 25.2595 - val_loss: 34.6033 - val_mean_absolute_error: 34.6033\n",
            "Epoch 29/200\n",
            "22/22 [==============================] - 1s 25ms/step - loss: 25.6302 - mean_absolute_error: 25.6302 - val_loss: 35.2006 - val_mean_absolute_error: 35.2006\n",
            "Epoch 30/200\n",
            "22/22 [==============================] - 1s 24ms/step - loss: 37.0984 - mean_absolute_error: 37.0984 - val_loss: 44.3377 - val_mean_absolute_error: 44.3377\n",
            "Epoch 31/200\n",
            "22/22 [==============================] - 1s 25ms/step - loss: 30.2743 - mean_absolute_error: 30.2743 - val_loss: 34.9425 - val_mean_absolute_error: 34.9425\n",
            "Epoch 32/200\n",
            "22/22 [==============================] - 1s 26ms/step - loss: 30.4113 - mean_absolute_error: 30.4113 - val_loss: 41.5521 - val_mean_absolute_error: 41.5521\n",
            "Epoch 33/200\n",
            "22/22 [==============================] - 1s 25ms/step - loss: 27.0059 - mean_absolute_error: 27.0059 - val_loss: 33.2145 - val_mean_absolute_error: 33.2145\n",
            "Epoch 34/200\n",
            "22/22 [==============================] - 1s 24ms/step - loss: 24.6841 - mean_absolute_error: 24.6841 - val_loss: 37.8025 - val_mean_absolute_error: 37.8025\n",
            "Epoch 35/200\n",
            "22/22 [==============================] - 1s 26ms/step - loss: 25.5229 - mean_absolute_error: 25.5229 - val_loss: 35.1191 - val_mean_absolute_error: 35.1191\n",
            "Epoch 36/200\n",
            "22/22 [==============================] - 1s 24ms/step - loss: 24.0446 - mean_absolute_error: 24.0446 - val_loss: 40.7069 - val_mean_absolute_error: 40.7069\n",
            "Epoch 37/200\n",
            "22/22 [==============================] - 1s 26ms/step - loss: 27.5150 - mean_absolute_error: 27.5150 - val_loss: 34.7909 - val_mean_absolute_error: 34.7909\n",
            "Epoch 38/200\n",
            "22/22 [==============================] - 1s 26ms/step - loss: 23.6377 - mean_absolute_error: 23.6377 - val_loss: 33.2661 - val_mean_absolute_error: 33.2661\n",
            "Epoch 39/200\n",
            "22/22 [==============================] - 1s 23ms/step - loss: 22.7839 - mean_absolute_error: 22.7839 - val_loss: 32.9149 - val_mean_absolute_error: 32.9149\n",
            "Epoch 40/200\n",
            "22/22 [==============================] - 1s 25ms/step - loss: 24.3521 - mean_absolute_error: 24.3521 - val_loss: 44.2089 - val_mean_absolute_error: 44.2089\n",
            "Epoch 41/200\n",
            "22/22 [==============================] - 1s 25ms/step - loss: 23.7922 - mean_absolute_error: 23.7922 - val_loss: 32.8271 - val_mean_absolute_error: 32.8271\n",
            "Epoch 42/200\n",
            "22/22 [==============================] - 1s 24ms/step - loss: 22.5681 - mean_absolute_error: 22.5681 - val_loss: 33.5627 - val_mean_absolute_error: 33.5627\n",
            "Epoch 43/200\n",
            "22/22 [==============================] - 1s 25ms/step - loss: 24.3958 - mean_absolute_error: 24.3958 - val_loss: 33.6756 - val_mean_absolute_error: 33.6756\n",
            "Epoch 44/200\n",
            "22/22 [==============================] - 1s 24ms/step - loss: 21.8621 - mean_absolute_error: 21.8621 - val_loss: 34.1505 - val_mean_absolute_error: 34.1505\n",
            "Epoch 45/200\n",
            "22/22 [==============================] - 1s 23ms/step - loss: 22.4138 - mean_absolute_error: 22.4138 - val_loss: 38.0731 - val_mean_absolute_error: 38.0731\n",
            "Epoch 46/200\n",
            "22/22 [==============================] - 1s 25ms/step - loss: 23.3578 - mean_absolute_error: 23.3578 - val_loss: 35.2340 - val_mean_absolute_error: 35.2340\n",
            "Epoch 47/200\n",
            "22/22 [==============================] - 1s 26ms/step - loss: 22.5659 - mean_absolute_error: 22.5659 - val_loss: 39.3373 - val_mean_absolute_error: 39.3373\n",
            "Epoch 48/200\n",
            "22/22 [==============================] - 1s 25ms/step - loss: 21.5554 - mean_absolute_error: 21.5554 - val_loss: 31.3362 - val_mean_absolute_error: 31.3362\n",
            "Epoch 49/200\n",
            "22/22 [==============================] - 1s 27ms/step - loss: 18.9379 - mean_absolute_error: 18.9379 - val_loss: 32.1720 - val_mean_absolute_error: 32.1720\n",
            "Epoch 50/200\n",
            "22/22 [==============================] - 1s 25ms/step - loss: 18.0583 - mean_absolute_error: 18.0583 - val_loss: 31.2303 - val_mean_absolute_error: 31.2303\n",
            "Epoch 51/200\n",
            "22/22 [==============================] - 1s 24ms/step - loss: 20.8819 - mean_absolute_error: 20.8819 - val_loss: 33.7849 - val_mean_absolute_error: 33.7849\n",
            "Epoch 52/200\n",
            "22/22 [==============================] - 1s 24ms/step - loss: 19.9810 - mean_absolute_error: 19.9810 - val_loss: 32.6075 - val_mean_absolute_error: 32.6075\n",
            "Epoch 53/200\n",
            "22/22 [==============================] - 1s 26ms/step - loss: 20.5026 - mean_absolute_error: 20.5026 - val_loss: 35.0823 - val_mean_absolute_error: 35.0823\n",
            "Epoch 54/200\n",
            "22/22 [==============================] - 1s 24ms/step - loss: 20.0904 - mean_absolute_error: 20.0904 - val_loss: 30.7389 - val_mean_absolute_error: 30.7389\n",
            "Epoch 55/200\n",
            "22/22 [==============================] - 1s 24ms/step - loss: 20.5925 - mean_absolute_error: 20.5925 - val_loss: 31.5248 - val_mean_absolute_error: 31.5248\n",
            "Epoch 56/200\n",
            "22/22 [==============================] - 1s 24ms/step - loss: 20.6779 - mean_absolute_error: 20.6779 - val_loss: 33.4858 - val_mean_absolute_error: 33.4858\n",
            "Epoch 57/200\n",
            "22/22 [==============================] - 1s 26ms/step - loss: 22.6273 - mean_absolute_error: 22.6273 - val_loss: 35.1878 - val_mean_absolute_error: 35.1878\n",
            "Epoch 58/200\n",
            "22/22 [==============================] - 1s 25ms/step - loss: 25.6622 - mean_absolute_error: 25.6622 - val_loss: 34.7963 - val_mean_absolute_error: 34.7963\n",
            "Epoch 59/200\n",
            "22/22 [==============================] - 1s 28ms/step - loss: 21.2759 - mean_absolute_error: 21.2759 - val_loss: 41.2899 - val_mean_absolute_error: 41.2899\n",
            "Epoch 60/200\n",
            "22/22 [==============================] - 1s 25ms/step - loss: 20.6645 - mean_absolute_error: 20.6645 - val_loss: 31.0390 - val_mean_absolute_error: 31.0390\n",
            "Epoch 61/200\n",
            "22/22 [==============================] - 1s 24ms/step - loss: 16.6017 - mean_absolute_error: 16.6017 - val_loss: 29.8139 - val_mean_absolute_error: 29.8139\n",
            "Epoch 62/200\n",
            "22/22 [==============================] - 1s 24ms/step - loss: 16.3648 - mean_absolute_error: 16.3648 - val_loss: 31.6486 - val_mean_absolute_error: 31.6486\n",
            "Epoch 63/200\n",
            "22/22 [==============================] - 1s 26ms/step - loss: 18.3098 - mean_absolute_error: 18.3098 - val_loss: 29.0559 - val_mean_absolute_error: 29.0559\n",
            "Epoch 64/200\n",
            "22/22 [==============================] - 1s 25ms/step - loss: 17.3781 - mean_absolute_error: 17.3781 - val_loss: 29.9713 - val_mean_absolute_error: 29.9713\n",
            "Epoch 65/200\n",
            "22/22 [==============================] - 1s 27ms/step - loss: 21.0186 - mean_absolute_error: 21.0186 - val_loss: 38.4418 - val_mean_absolute_error: 38.4418\n",
            "Epoch 66/200\n",
            "22/22 [==============================] - 1s 26ms/step - loss: 20.2387 - mean_absolute_error: 20.2387 - val_loss: 32.3587 - val_mean_absolute_error: 32.3587\n",
            "Epoch 67/200\n",
            "22/22 [==============================] - 1s 24ms/step - loss: 15.9267 - mean_absolute_error: 15.9267 - val_loss: 28.3921 - val_mean_absolute_error: 28.3921\n",
            "Epoch 68/200\n",
            "22/22 [==============================] - 1s 27ms/step - loss: 14.6949 - mean_absolute_error: 14.6949 - val_loss: 29.0469 - val_mean_absolute_error: 29.0469\n",
            "Epoch 69/200\n",
            "22/22 [==============================] - 1s 25ms/step - loss: 16.5953 - mean_absolute_error: 16.5953 - val_loss: 29.7056 - val_mean_absolute_error: 29.7056\n",
            "Epoch 70/200\n",
            "22/22 [==============================] - 1s 26ms/step - loss: 14.3623 - mean_absolute_error: 14.3623 - val_loss: 29.6018 - val_mean_absolute_error: 29.6018\n",
            "Epoch 71/200\n",
            "22/22 [==============================] - 1s 26ms/step - loss: 15.3904 - mean_absolute_error: 15.3904 - val_loss: 29.9956 - val_mean_absolute_error: 29.9956\n",
            "Epoch 72/200\n",
            "22/22 [==============================] - 1s 25ms/step - loss: 19.5532 - mean_absolute_error: 19.5532 - val_loss: 31.9671 - val_mean_absolute_error: 31.9671\n",
            "Epoch 73/200\n",
            "22/22 [==============================] - 1s 24ms/step - loss: 17.0485 - mean_absolute_error: 17.0485 - val_loss: 30.8340 - val_mean_absolute_error: 30.8340\n",
            "Epoch 74/200\n",
            "22/22 [==============================] - 1s 27ms/step - loss: 17.6231 - mean_absolute_error: 17.6231 - val_loss: 29.9346 - val_mean_absolute_error: 29.9346\n",
            "Epoch 75/200\n",
            "22/22 [==============================] - 1s 28ms/step - loss: 24.6181 - mean_absolute_error: 24.6181 - val_loss: 33.1036 - val_mean_absolute_error: 33.1036\n",
            "Epoch 76/200\n",
            "22/22 [==============================] - 1s 25ms/step - loss: 24.9596 - mean_absolute_error: 24.9596 - val_loss: 31.1320 - val_mean_absolute_error: 31.1320\n",
            "Epoch 77/200\n",
            "22/22 [==============================] - 1s 27ms/step - loss: 21.4862 - mean_absolute_error: 21.4862 - val_loss: 29.4106 - val_mean_absolute_error: 29.4106\n",
            "Epoch 78/200\n",
            "22/22 [==============================] - 1s 26ms/step - loss: 17.5735 - mean_absolute_error: 17.5735 - val_loss: 29.7154 - val_mean_absolute_error: 29.7154\n",
            "Epoch 79/200\n",
            "22/22 [==============================] - 1s 26ms/step - loss: 15.9823 - mean_absolute_error: 15.9823 - val_loss: 27.5503 - val_mean_absolute_error: 27.5503\n",
            "Epoch 80/200\n",
            "22/22 [==============================] - 1s 27ms/step - loss: 14.0570 - mean_absolute_error: 14.0570 - val_loss: 29.2291 - val_mean_absolute_error: 29.2291\n",
            "Epoch 81/200\n",
            "22/22 [==============================] - 1s 26ms/step - loss: 16.3094 - mean_absolute_error: 16.3094 - val_loss: 28.2941 - val_mean_absolute_error: 28.2941\n",
            "Epoch 82/200\n",
            "22/22 [==============================] - 1s 27ms/step - loss: 16.0645 - mean_absolute_error: 16.0645 - val_loss: 26.7564 - val_mean_absolute_error: 26.7564\n",
            "Epoch 83/200\n",
            "22/22 [==============================] - 1s 26ms/step - loss: 14.7356 - mean_absolute_error: 14.7356 - val_loss: 31.5218 - val_mean_absolute_error: 31.5218\n",
            "Epoch 84/200\n",
            "22/22 [==============================] - 1s 26ms/step - loss: 15.1176 - mean_absolute_error: 15.1176 - val_loss: 27.8563 - val_mean_absolute_error: 27.8563\n",
            "Epoch 85/200\n",
            "22/22 [==============================] - 1s 26ms/step - loss: 13.5133 - mean_absolute_error: 13.5133 - val_loss: 27.7073 - val_mean_absolute_error: 27.7073\n",
            "Epoch 86/200\n",
            "22/22 [==============================] - 1s 25ms/step - loss: 14.6564 - mean_absolute_error: 14.6564 - val_loss: 30.5428 - val_mean_absolute_error: 30.5428\n",
            "Epoch 87/200\n",
            "22/22 [==============================] - 1s 25ms/step - loss: 18.6356 - mean_absolute_error: 18.6356 - val_loss: 28.3961 - val_mean_absolute_error: 28.3961\n",
            "Epoch 88/200\n",
            "22/22 [==============================] - 1s 26ms/step - loss: 15.8413 - mean_absolute_error: 15.8413 - val_loss: 28.4076 - val_mean_absolute_error: 28.4076\n",
            "Epoch 89/200\n",
            "22/22 [==============================] - 1s 25ms/step - loss: 13.2725 - mean_absolute_error: 13.2725 - val_loss: 30.9278 - val_mean_absolute_error: 30.9278\n",
            "Epoch 90/200\n",
            "22/22 [==============================] - 1s 24ms/step - loss: 16.1325 - mean_absolute_error: 16.1325 - val_loss: 28.0566 - val_mean_absolute_error: 28.0566\n",
            "Epoch 91/200\n",
            "22/22 [==============================] - 1s 25ms/step - loss: 13.9901 - mean_absolute_error: 13.9901 - val_loss: 26.2024 - val_mean_absolute_error: 26.2024\n",
            "Epoch 92/200\n",
            "22/22 [==============================] - 1s 25ms/step - loss: 18.1354 - mean_absolute_error: 18.1354 - val_loss: 42.5170 - val_mean_absolute_error: 42.5170\n",
            "Epoch 93/200\n",
            "22/22 [==============================] - 1s 24ms/step - loss: 20.2159 - mean_absolute_error: 20.2159 - val_loss: 28.8777 - val_mean_absolute_error: 28.8777\n",
            "Epoch 94/200\n",
            "22/22 [==============================] - 1s 25ms/step - loss: 21.3391 - mean_absolute_error: 21.3391 - val_loss: 32.8120 - val_mean_absolute_error: 32.8120\n",
            "Epoch 95/200\n",
            "22/22 [==============================] - 1s 27ms/step - loss: 18.7455 - mean_absolute_error: 18.7455 - val_loss: 27.1615 - val_mean_absolute_error: 27.1615\n",
            "Epoch 96/200\n",
            "22/22 [==============================] - 1s 24ms/step - loss: 15.6805 - mean_absolute_error: 15.6805 - val_loss: 30.5905 - val_mean_absolute_error: 30.5905\n",
            "Epoch 97/200\n",
            "22/22 [==============================] - 1s 25ms/step - loss: 17.8398 - mean_absolute_error: 17.8398 - val_loss: 29.4997 - val_mean_absolute_error: 29.4997\n",
            "Epoch 98/200\n",
            "22/22 [==============================] - 1s 26ms/step - loss: 18.2970 - mean_absolute_error: 18.2970 - val_loss: 29.3500 - val_mean_absolute_error: 29.3500\n",
            "Epoch 99/200\n",
            "22/22 [==============================] - 1s 26ms/step - loss: 14.2979 - mean_absolute_error: 14.2979 - val_loss: 26.8894 - val_mean_absolute_error: 26.8894\n",
            "Epoch 100/200\n",
            "22/22 [==============================] - 1s 25ms/step - loss: 12.2155 - mean_absolute_error: 12.2155 - val_loss: 28.5258 - val_mean_absolute_error: 28.5258\n",
            "Epoch 101/200\n",
            "22/22 [==============================] - 1s 25ms/step - loss: 11.0936 - mean_absolute_error: 11.0936 - val_loss: 31.9540 - val_mean_absolute_error: 31.9540\n",
            "Epoch 102/200\n",
            "22/22 [==============================] - 1s 27ms/step - loss: 14.7193 - mean_absolute_error: 14.7193 - val_loss: 34.8228 - val_mean_absolute_error: 34.8228\n",
            "Epoch 103/200\n",
            "22/22 [==============================] - 1s 25ms/step - loss: 14.1043 - mean_absolute_error: 14.1043 - val_loss: 30.0831 - val_mean_absolute_error: 30.0831\n",
            "Epoch 104/200\n",
            "22/22 [==============================] - 1s 25ms/step - loss: 12.0771 - mean_absolute_error: 12.0771 - val_loss: 27.2210 - val_mean_absolute_error: 27.2210\n",
            "Epoch 105/200\n",
            "22/22 [==============================] - 1s 25ms/step - loss: 12.7181 - mean_absolute_error: 12.7181 - val_loss: 30.9642 - val_mean_absolute_error: 30.9642\n",
            "Epoch 106/200\n",
            "22/22 [==============================] - 1s 25ms/step - loss: 13.9992 - mean_absolute_error: 13.9992 - val_loss: 27.9103 - val_mean_absolute_error: 27.9103\n",
            "Epoch 107/200\n",
            "22/22 [==============================] - 1s 26ms/step - loss: 12.7247 - mean_absolute_error: 12.7247 - val_loss: 30.9634 - val_mean_absolute_error: 30.9634\n",
            "Epoch 108/200\n",
            "22/22 [==============================] - 1s 25ms/step - loss: 13.9418 - mean_absolute_error: 13.9418 - val_loss: 30.9841 - val_mean_absolute_error: 30.9841\n",
            "Epoch 109/200\n",
            "22/22 [==============================] - 1s 25ms/step - loss: 16.2266 - mean_absolute_error: 16.2266 - val_loss: 26.5071 - val_mean_absolute_error: 26.5071\n",
            "Epoch 110/200\n",
            "22/22 [==============================] - 1s 24ms/step - loss: 12.3327 - mean_absolute_error: 12.3327 - val_loss: 25.2844 - val_mean_absolute_error: 25.2844\n",
            "Epoch 111/200\n",
            "22/22 [==============================] - 1s 25ms/step - loss: 12.8529 - mean_absolute_error: 12.8529 - val_loss: 27.0804 - val_mean_absolute_error: 27.0804\n",
            "Epoch 112/200\n",
            "22/22 [==============================] - 1s 27ms/step - loss: 9.8803 - mean_absolute_error: 9.8803 - val_loss: 25.0938 - val_mean_absolute_error: 25.0938\n",
            "Epoch 113/200\n",
            "22/22 [==============================] - 1s 26ms/step - loss: 10.4318 - mean_absolute_error: 10.4318 - val_loss: 29.1015 - val_mean_absolute_error: 29.1015\n",
            "Epoch 114/200\n",
            "22/22 [==============================] - 1s 25ms/step - loss: 10.4176 - mean_absolute_error: 10.4176 - val_loss: 26.1147 - val_mean_absolute_error: 26.1147\n",
            "Epoch 115/200\n",
            "22/22 [==============================] - 1s 25ms/step - loss: 12.3601 - mean_absolute_error: 12.3601 - val_loss: 26.4177 - val_mean_absolute_error: 26.4177\n",
            "Epoch 116/200\n",
            "22/22 [==============================] - 1s 26ms/step - loss: 11.6400 - mean_absolute_error: 11.6400 - val_loss: 27.2882 - val_mean_absolute_error: 27.2882\n",
            "Epoch 117/200\n",
            "22/22 [==============================] - 1s 25ms/step - loss: 11.1881 - mean_absolute_error: 11.1881 - val_loss: 28.1088 - val_mean_absolute_error: 28.1088\n",
            "Epoch 118/200\n",
            "22/22 [==============================] - 1s 25ms/step - loss: 11.5758 - mean_absolute_error: 11.5758 - val_loss: 25.6091 - val_mean_absolute_error: 25.6091\n",
            "Epoch 119/200\n",
            "22/22 [==============================] - 1s 26ms/step - loss: 10.6301 - mean_absolute_error: 10.6301 - val_loss: 26.8730 - val_mean_absolute_error: 26.8730\n",
            "Epoch 120/200\n",
            "22/22 [==============================] - 1s 24ms/step - loss: 12.2542 - mean_absolute_error: 12.2542 - val_loss: 29.6082 - val_mean_absolute_error: 29.6082\n",
            "Epoch 121/200\n",
            "22/22 [==============================] - 1s 26ms/step - loss: 11.9060 - mean_absolute_error: 11.9060 - val_loss: 27.8780 - val_mean_absolute_error: 27.8780\n",
            "Epoch 122/200\n",
            "22/22 [==============================] - 1s 25ms/step - loss: 11.3511 - mean_absolute_error: 11.3511 - val_loss: 27.2117 - val_mean_absolute_error: 27.2117\n",
            "Epoch 123/200\n",
            "22/22 [==============================] - 1s 25ms/step - loss: 11.7048 - mean_absolute_error: 11.7048 - val_loss: 29.9197 - val_mean_absolute_error: 29.9197\n",
            "Epoch 124/200\n",
            "22/22 [==============================] - 1s 26ms/step - loss: 10.5001 - mean_absolute_error: 10.5001 - val_loss: 25.3178 - val_mean_absolute_error: 25.3178\n",
            "Epoch 125/200\n",
            "22/22 [==============================] - 1s 25ms/step - loss: 14.7565 - mean_absolute_error: 14.7565 - val_loss: 33.2484 - val_mean_absolute_error: 33.2484\n",
            "Epoch 126/200\n",
            "22/22 [==============================] - 1s 25ms/step - loss: 18.0407 - mean_absolute_error: 18.0407 - val_loss: 27.5045 - val_mean_absolute_error: 27.5045\n",
            "Epoch 127/200\n",
            "22/22 [==============================] - 1s 25ms/step - loss: 11.3709 - mean_absolute_error: 11.3709 - val_loss: 26.2309 - val_mean_absolute_error: 26.2309\n",
            "Epoch 128/200\n",
            "22/22 [==============================] - 1s 25ms/step - loss: 10.2363 - mean_absolute_error: 10.2363 - val_loss: 27.2325 - val_mean_absolute_error: 27.2325\n",
            "Epoch 129/200\n",
            "22/22 [==============================] - 1s 24ms/step - loss: 9.4096 - mean_absolute_error: 9.4096 - val_loss: 25.3831 - val_mean_absolute_error: 25.3831\n",
            "Epoch 130/200\n",
            "22/22 [==============================] - 1s 25ms/step - loss: 9.5643 - mean_absolute_error: 9.5643 - val_loss: 26.1881 - val_mean_absolute_error: 26.1881\n",
            "Epoch 131/200\n",
            "22/22 [==============================] - 1s 25ms/step - loss: 7.9318 - mean_absolute_error: 7.9318 - val_loss: 26.2172 - val_mean_absolute_error: 26.2172\n",
            "Epoch 132/200\n",
            "22/22 [==============================] - 1s 27ms/step - loss: 12.9523 - mean_absolute_error: 12.9523 - val_loss: 32.1627 - val_mean_absolute_error: 32.1627\n",
            "Epoch 133/200\n",
            "22/22 [==============================] - 1s 27ms/step - loss: 11.2431 - mean_absolute_error: 11.2431 - val_loss: 27.1030 - val_mean_absolute_error: 27.1030\n",
            "Epoch 134/200\n",
            "22/22 [==============================] - 1s 25ms/step - loss: 10.9865 - mean_absolute_error: 10.9865 - val_loss: 27.1627 - val_mean_absolute_error: 27.1627\n",
            "Epoch 135/200\n",
            "22/22 [==============================] - 1s 25ms/step - loss: 11.5818 - mean_absolute_error: 11.5818 - val_loss: 27.6720 - val_mean_absolute_error: 27.6720\n",
            "Epoch 136/200\n",
            "22/22 [==============================] - 1s 25ms/step - loss: 9.1342 - mean_absolute_error: 9.1342 - val_loss: 28.3173 - val_mean_absolute_error: 28.3173\n",
            "Epoch 137/200\n",
            "22/22 [==============================] - 1s 24ms/step - loss: 9.0649 - mean_absolute_error: 9.0649 - val_loss: 25.6382 - val_mean_absolute_error: 25.6382\n",
            "Epoch 138/200\n",
            "22/22 [==============================] - 1s 25ms/step - loss: 8.8437 - mean_absolute_error: 8.8437 - val_loss: 28.2128 - val_mean_absolute_error: 28.2128\n",
            "Epoch 139/200\n",
            "22/22 [==============================] - 1s 24ms/step - loss: 12.2933 - mean_absolute_error: 12.2933 - val_loss: 28.5083 - val_mean_absolute_error: 28.5083\n",
            "Epoch 140/200\n",
            "22/22 [==============================] - 1s 26ms/step - loss: 10.2487 - mean_absolute_error: 10.2487 - val_loss: 27.7495 - val_mean_absolute_error: 27.7495\n",
            "Epoch 141/200\n",
            "22/22 [==============================] - 1s 25ms/step - loss: 9.4142 - mean_absolute_error: 9.4142 - val_loss: 27.5785 - val_mean_absolute_error: 27.5785\n",
            "Epoch 142/200\n",
            "22/22 [==============================] - 1s 25ms/step - loss: 7.4227 - mean_absolute_error: 7.4227 - val_loss: 26.6277 - val_mean_absolute_error: 26.6277\n",
            "Epoch 143/200\n",
            "22/22 [==============================] - 1s 27ms/step - loss: 7.9994 - mean_absolute_error: 7.9994 - val_loss: 26.3883 - val_mean_absolute_error: 26.3883\n",
            "Epoch 144/200\n",
            "22/22 [==============================] - 1s 25ms/step - loss: 11.1774 - mean_absolute_error: 11.1774 - val_loss: 28.2908 - val_mean_absolute_error: 28.2908\n",
            "Epoch 145/200\n",
            "22/22 [==============================] - 1s 28ms/step - loss: 8.4438 - mean_absolute_error: 8.4438 - val_loss: 25.6871 - val_mean_absolute_error: 25.6871\n",
            "Epoch 146/200\n",
            "22/22 [==============================] - 1s 25ms/step - loss: 9.0796 - mean_absolute_error: 9.0796 - val_loss: 25.4231 - val_mean_absolute_error: 25.4231\n",
            "Epoch 147/200\n",
            "22/22 [==============================] - 1s 25ms/step - loss: 8.3878 - mean_absolute_error: 8.3878 - val_loss: 26.2026 - val_mean_absolute_error: 26.2026\n",
            "Epoch 148/200\n",
            "22/22 [==============================] - 1s 24ms/step - loss: 8.3704 - mean_absolute_error: 8.3704 - val_loss: 28.2400 - val_mean_absolute_error: 28.2400\n",
            "Epoch 149/200\n",
            "22/22 [==============================] - 1s 27ms/step - loss: 9.1938 - mean_absolute_error: 9.1938 - val_loss: 30.4096 - val_mean_absolute_error: 30.4096\n",
            "Epoch 150/200\n",
            "22/22 [==============================] - 1s 24ms/step - loss: 9.8189 - mean_absolute_error: 9.8189 - val_loss: 28.4453 - val_mean_absolute_error: 28.4453\n",
            "Epoch 151/200\n",
            "22/22 [==============================] - 1s 26ms/step - loss: 9.6148 - mean_absolute_error: 9.6148 - val_loss: 27.1415 - val_mean_absolute_error: 27.1415\n",
            "Epoch 152/200\n",
            "22/22 [==============================] - 1s 25ms/step - loss: 9.3412 - mean_absolute_error: 9.3412 - val_loss: 25.8059 - val_mean_absolute_error: 25.8059\n",
            "Epoch 153/200\n",
            "22/22 [==============================] - 1s 25ms/step - loss: 9.0381 - mean_absolute_error: 9.0381 - val_loss: 28.9358 - val_mean_absolute_error: 28.9358\n",
            "Epoch 154/200\n",
            "22/22 [==============================] - 1s 26ms/step - loss: 12.5623 - mean_absolute_error: 12.5623 - val_loss: 28.4575 - val_mean_absolute_error: 28.4575\n",
            "Epoch 155/200\n",
            "22/22 [==============================] - 1s 27ms/step - loss: 9.2855 - mean_absolute_error: 9.2855 - val_loss: 26.4309 - val_mean_absolute_error: 26.4309\n",
            "Epoch 156/200\n",
            "22/22 [==============================] - 1s 26ms/step - loss: 8.7311 - mean_absolute_error: 8.7311 - val_loss: 26.9868 - val_mean_absolute_error: 26.9868\n",
            "Epoch 157/200\n",
            "22/22 [==============================] - 1s 24ms/step - loss: 9.8034 - mean_absolute_error: 9.8034 - val_loss: 28.8299 - val_mean_absolute_error: 28.8299\n",
            "Epoch 158/200\n",
            "22/22 [==============================] - 1s 26ms/step - loss: 7.5686 - mean_absolute_error: 7.5686 - val_loss: 27.5973 - val_mean_absolute_error: 27.5973\n",
            "Epoch 159/200\n",
            "22/22 [==============================] - 1s 25ms/step - loss: 9.5132 - mean_absolute_error: 9.5132 - val_loss: 28.4710 - val_mean_absolute_error: 28.4710\n",
            "Epoch 160/200\n",
            "22/22 [==============================] - 1s 25ms/step - loss: 10.5933 - mean_absolute_error: 10.5933 - val_loss: 25.8797 - val_mean_absolute_error: 25.8797\n",
            "Epoch 161/200\n",
            "22/22 [==============================] - 1s 25ms/step - loss: 6.8703 - mean_absolute_error: 6.8703 - val_loss: 28.4432 - val_mean_absolute_error: 28.4432\n",
            "Epoch 162/200\n",
            "22/22 [==============================] - 1s 25ms/step - loss: 7.4404 - mean_absolute_error: 7.4404 - val_loss: 30.8995 - val_mean_absolute_error: 30.8995\n",
            "Epoch 163/200\n",
            "22/22 [==============================] - 1s 27ms/step - loss: 10.0906 - mean_absolute_error: 10.0906 - val_loss: 27.2315 - val_mean_absolute_error: 27.2315\n",
            "Epoch 164/200\n",
            "22/22 [==============================] - 1s 25ms/step - loss: 9.9198 - mean_absolute_error: 9.9198 - val_loss: 34.1290 - val_mean_absolute_error: 34.1290\n",
            "Epoch 165/200\n",
            "22/22 [==============================] - 1s 26ms/step - loss: 14.3236 - mean_absolute_error: 14.3236 - val_loss: 27.2982 - val_mean_absolute_error: 27.2982\n",
            "Epoch 166/200\n",
            "22/22 [==============================] - 1s 25ms/step - loss: 8.2276 - mean_absolute_error: 8.2276 - val_loss: 27.8492 - val_mean_absolute_error: 27.8492\n",
            "Epoch 167/200\n",
            "22/22 [==============================] - 1s 27ms/step - loss: 7.6048 - mean_absolute_error: 7.6048 - val_loss: 26.2859 - val_mean_absolute_error: 26.2859\n",
            "Epoch 168/200\n",
            "22/22 [==============================] - 1s 25ms/step - loss: 7.1365 - mean_absolute_error: 7.1365 - val_loss: 27.4653 - val_mean_absolute_error: 27.4653\n",
            "Epoch 169/200\n",
            "22/22 [==============================] - 1s 26ms/step - loss: 7.5865 - mean_absolute_error: 7.5865 - val_loss: 26.3120 - val_mean_absolute_error: 26.3120\n",
            "Epoch 170/200\n",
            "22/22 [==============================] - 1s 28ms/step - loss: 7.6833 - mean_absolute_error: 7.6833 - val_loss: 27.6966 - val_mean_absolute_error: 27.6966\n",
            "Epoch 171/200\n",
            "22/22 [==============================] - 1s 25ms/step - loss: 9.4763 - mean_absolute_error: 9.4763 - val_loss: 27.0952 - val_mean_absolute_error: 27.0952\n",
            "Epoch 172/200\n",
            "22/22 [==============================] - 1s 25ms/step - loss: 8.4946 - mean_absolute_error: 8.4946 - val_loss: 29.2635 - val_mean_absolute_error: 29.2635\n",
            "Epoch 173/200\n",
            "22/22 [==============================] - 1s 25ms/step - loss: 7.4620 - mean_absolute_error: 7.4620 - val_loss: 26.0970 - val_mean_absolute_error: 26.0970\n",
            "Epoch 174/200\n",
            "22/22 [==============================] - 1s 27ms/step - loss: 7.7579 - mean_absolute_error: 7.7579 - val_loss: 27.2285 - val_mean_absolute_error: 27.2285\n",
            "Epoch 175/200\n",
            "22/22 [==============================] - 1s 26ms/step - loss: 8.4692 - mean_absolute_error: 8.4692 - val_loss: 28.5871 - val_mean_absolute_error: 28.5871\n",
            "Epoch 176/200\n",
            "22/22 [==============================] - 1s 26ms/step - loss: 8.7810 - mean_absolute_error: 8.7810 - val_loss: 27.8079 - val_mean_absolute_error: 27.8079\n",
            "Epoch 177/200\n",
            "22/22 [==============================] - 1s 25ms/step - loss: 6.6373 - mean_absolute_error: 6.6373 - val_loss: 27.4808 - val_mean_absolute_error: 27.4808\n",
            "Epoch 178/200\n",
            "22/22 [==============================] - 1s 25ms/step - loss: 7.9411 - mean_absolute_error: 7.9411 - val_loss: 28.6404 - val_mean_absolute_error: 28.6404\n",
            "Epoch 179/200\n",
            "22/22 [==============================] - 1s 26ms/step - loss: 6.6666 - mean_absolute_error: 6.6666 - val_loss: 28.1789 - val_mean_absolute_error: 28.1789\n",
            "Epoch 180/200\n",
            "22/22 [==============================] - 1s 27ms/step - loss: 6.5019 - mean_absolute_error: 6.5019 - val_loss: 27.1824 - val_mean_absolute_error: 27.1824\n",
            "Epoch 181/200\n",
            "22/22 [==============================] - 1s 26ms/step - loss: 6.0562 - mean_absolute_error: 6.0562 - val_loss: 26.4142 - val_mean_absolute_error: 26.4142\n",
            "Epoch 182/200\n",
            "22/22 [==============================] - 1s 25ms/step - loss: 5.9208 - mean_absolute_error: 5.9208 - val_loss: 28.4301 - val_mean_absolute_error: 28.4301\n",
            "Epoch 183/200\n",
            "22/22 [==============================] - 1s 26ms/step - loss: 6.9802 - mean_absolute_error: 6.9802 - val_loss: 25.7804 - val_mean_absolute_error: 25.7804\n",
            "Epoch 184/200\n",
            "22/22 [==============================] - 1s 27ms/step - loss: 6.5701 - mean_absolute_error: 6.5701 - val_loss: 29.1938 - val_mean_absolute_error: 29.1938\n",
            "Epoch 185/200\n",
            "22/22 [==============================] - 1s 27ms/step - loss: 8.6454 - mean_absolute_error: 8.6454 - val_loss: 25.7716 - val_mean_absolute_error: 25.7716\n",
            "Epoch 186/200\n",
            "22/22 [==============================] - 1s 27ms/step - loss: 6.1468 - mean_absolute_error: 6.1468 - val_loss: 26.8469 - val_mean_absolute_error: 26.8469\n",
            "Epoch 187/200\n",
            "22/22 [==============================] - 1s 26ms/step - loss: 7.4820 - mean_absolute_error: 7.4820 - val_loss: 29.7709 - val_mean_absolute_error: 29.7709\n",
            "Epoch 188/200\n",
            "22/22 [==============================] - 1s 26ms/step - loss: 7.8085 - mean_absolute_error: 7.8085 - val_loss: 27.4349 - val_mean_absolute_error: 27.4349\n",
            "Epoch 189/200\n",
            "22/22 [==============================] - 1s 27ms/step - loss: 7.3364 - mean_absolute_error: 7.3364 - val_loss: 31.0119 - val_mean_absolute_error: 31.0119\n",
            "Epoch 190/200\n",
            "22/22 [==============================] - 1s 26ms/step - loss: 10.1340 - mean_absolute_error: 10.1340 - val_loss: 28.4609 - val_mean_absolute_error: 28.4609\n",
            "Epoch 191/200\n",
            "22/22 [==============================] - 1s 26ms/step - loss: 8.3322 - mean_absolute_error: 8.3322 - val_loss: 27.0916 - val_mean_absolute_error: 27.0916\n",
            "Epoch 192/200\n",
            "22/22 [==============================] - 1s 26ms/step - loss: 7.6003 - mean_absolute_error: 7.6003 - val_loss: 26.2596 - val_mean_absolute_error: 26.2596\n",
            "Epoch 193/200\n",
            "22/22 [==============================] - 1s 26ms/step - loss: 6.9961 - mean_absolute_error: 6.9961 - val_loss: 28.8615 - val_mean_absolute_error: 28.8615\n",
            "Epoch 194/200\n",
            "22/22 [==============================] - 1s 26ms/step - loss: 7.4303 - mean_absolute_error: 7.4303 - val_loss: 26.4659 - val_mean_absolute_error: 26.4659\n",
            "Epoch 195/200\n",
            "22/22 [==============================] - 1s 27ms/step - loss: 7.5525 - mean_absolute_error: 7.5525 - val_loss: 28.5732 - val_mean_absolute_error: 28.5732\n",
            "Epoch 196/200\n",
            "22/22 [==============================] - 1s 26ms/step - loss: 9.6743 - mean_absolute_error: 9.6743 - val_loss: 28.4024 - val_mean_absolute_error: 28.4024\n",
            "Epoch 197/200\n",
            "22/22 [==============================] - 1s 27ms/step - loss: 6.5790 - mean_absolute_error: 6.5790 - val_loss: 26.1934 - val_mean_absolute_error: 26.1934\n",
            "Epoch 198/200\n",
            "22/22 [==============================] - 1s 27ms/step - loss: 7.8886 - mean_absolute_error: 7.8886 - val_loss: 27.0975 - val_mean_absolute_error: 27.0975\n",
            "Epoch 199/200\n",
            "22/22 [==============================] - 1s 27ms/step - loss: 7.8583 - mean_absolute_error: 7.8583 - val_loss: 29.0865 - val_mean_absolute_error: 29.0865\n",
            "Epoch 200/200\n",
            "22/22 [==============================] - 1s 27ms/step - loss: 10.7305 - mean_absolute_error: 10.7305 - val_loss: 27.8552 - val_mean_absolute_error: 27.8552\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "MOxySfwnke8L"
      },
      "source": [
        "y_pred=best_model.predict(X_test)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "xB939UIJknjL",
        "outputId": "41831b5b-eae0-458c-946a-84bdbc319484"
      },
      "source": [
        "from sklearn.metrics import mean_absolute_error\n",
        "mean_absolute_error(y_test, y_pred)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "19.65605088462387"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 19
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "VOnaH4Tukmb7",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "fc795b07-25fa-414c-c279-600442a2e19d"
      },
      "source": [
        "print(model_hist.history.keys())\n"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "dict_keys(['loss', 'mean_absolute_error', 'val_loss', 'val_mean_absolute_error'])\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 295
        },
        "id": "LAV1BxU0oR_d",
        "outputId": "28364d9d-5647-4ac8-b917-7996c1c96bbc"
      },
      "source": [
        "import matplotlib.pyplot as plt\n",
        "plt.plot(model_hist.history['mean_absolute_error'])\n",
        "plt.plot(model_hist.history['val_mean_absolute_error'])\n",
        "plt.title('model mse')\n",
        "plt.ylabel('mse')\n",
        "plt.xlabel('epoch')\n",
        "plt.legend(['train', 'test'], loc='upper left')\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 295
        },
        "id": "XfFMyqVboai7",
        "outputId": "0b49bae2-6e4d-4aaf-9823-c159758953aa"
      },
      "source": [
        "#Loss function\n",
        "plt.plot(model_hist.history['loss'])\n",
        "plt.plot(model_hist.history['val_loss'])\n",
        "plt.title('model loss')\n",
        "plt.ylabel('loss')\n",
        "plt.xlabel('epoch')\n",
        "plt.legend(['train', 'test'], loc='upper left')\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ejxZ2Y_6o7ze"
      },
      "source": [
        "#loss decreases until the 100 epochs and then starts increasing for test_set"
      ],
      "execution_count": null,
      "outputs": []
    }
  ]
}