04_NN_Regresija_Regularizacija.ipynb

{
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/duskovesic/Machine-Learning/blob/main/NN_Regresija_Regularizacija.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "#Uvod"
      ],
      "metadata": {
        "id": "t6_oCQMAqBfJ"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "Kao i u prethodnoj svesci i u ovoj je primer rešavanja regresionog problema. Koristi se baza podataka „Boston Housing“. Ovde su kreirane četiri neuronske mreže. Prva mreža je ista kao u prethodnoj svesci. Prilikom treniranja druge mreže primenjena je regularizacija upotrebom L2 regularizatora kako bi se sprečila preprilagođenost modela podacima. U trećoj mreži su dodata dva nova „Dropout“ sloja, koja takođe sprečavaju preprilagodjenost. U četvrtoj mreži je upotrebljen i Dropout i L2 regularizator. Prilikom treniranja sva četiri modela, primenjena je „EarlyStopping“ tehnika koja prekida proces treniranja."
      ],
      "metadata": {
        "id": "gOjv0jw4qEFh"
      }
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "OG4k4oZGNcTV"
      },
      "source": [
        "# Importovanje biblioteka i setapovanje"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "e7gX1uHOs3bz"
      },
      "outputs": [],
      "source": [
        "import numpy as np\n",
        "\n",
        "from sklearn.preprocessing import StandardScaler\n",
        "from sklearn.metrics import r2_score\n",
        "\n",
        "from tensorflow.keras.datasets import boston_housing\n",
        "from tensorflow.keras.models import Sequential\n",
        "from tensorflow.keras.layers import InputLayer, Dense, Dropout\n",
        "from tensorflow.keras.optimizers import Adam\n",
        "from tensorflow.keras.losses import MeanSquaredError, MeanAbsoluteError\n",
        "from tensorflow.keras.callbacks import EarlyStopping\n",
        "from tensorflow.keras import regularizers\n",
        "\n",
        "from matplotlib import pyplot as plt\n",
        "\n",
        "RS = 123\n",
        "NUM_NEURONS_1 = 200\n",
        "NUM_NEURONS_2 = 100\n",
        "BATCH_SIZE = 64\n",
        "NUM_EPOCHS = 400"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "frxRzzUKNjBq"
      },
      "source": [
        "# Učitavanje podataka"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ildt6HvhuzzJ",
        "outputId": "080aa4ae-c75c-4dd9-b1b9-f5158614e66b"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Downloading data from https://storage.googleapis.com/tensorflow/tf-keras-datasets/boston_housing.npz\n",
            "57026/57026 [==============================] - 0s 0us/step\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "((339, 13), (167, 13), (339,), (167,))"
            ]
          },
          "metadata": {},
          "execution_count": 2
        }
      ],
      "source": [
        "(X_train, y_train), (X_test, y_test) = boston_housing.load_data(test_split=0.33, seed=RS)\n",
        "X_train.shape, X_test.shape, y_train.shape, y_test.shape"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "JJTNI6DoRxY0"
      },
      "source": [
        "# Standardizacija"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "DJwEVYWwSWVL",
        "outputId": "c8dec951-2287-4984-d593-2042fd1f7108"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(array([[5.11358e+01, 0.00000e+00, 1.81000e+01, ..., 2.02000e+01,\n",
              "         2.60000e+00, 1.01100e+01],\n",
              "        [5.73500e-02, 0.00000e+00, 4.49000e+00, ..., 1.85000e+01,\n",
              "         3.92300e+02, 6.53000e+00],\n",
              "        [3.57800e-02, 2.00000e+01, 3.33000e+00, ..., 1.49000e+01,\n",
              "         3.87310e+02, 3.76000e+00],\n",
              "        ...,\n",
              "        [9.51363e+00, 0.00000e+00, 1.81000e+01, ..., 2.02000e+01,\n",
              "         6.68000e+00, 1.87100e+01],\n",
              "        [2.44953e+00, 0.00000e+00, 1.95800e+01, ..., 1.47000e+01,\n",
              "         3.30040e+02, 1.13200e+01],\n",
              "        [1.96570e-01, 2.20000e+01, 5.86000e+00, ..., 1.91000e+01,\n",
              "         3.76140e+02, 1.01500e+01]]),\n",
              " array([[6.11540e-01, 2.00000e+01, 3.97000e+00, ..., 1.30000e+01,\n",
              "         3.89700e+02, 5.12000e+00],\n",
              "        [2.14090e-01, 2.20000e+01, 5.86000e+00, ..., 1.91000e+01,\n",
              "         3.77070e+02, 3.59000e+00],\n",
              "        [9.06500e-02, 2.00000e+01, 6.96000e+00, ..., 1.86000e+01,\n",
              "         3.91340e+02, 1.36500e+01],\n",
              "        ...,\n",
              "        [3.51140e-01, 0.00000e+00, 7.38000e+00, ..., 1.96000e+01,\n",
              "         3.96900e+02, 7.70000e+00],\n",
              "        [9.18702e+00, 0.00000e+00, 1.81000e+01, ..., 2.02000e+01,\n",
              "         3.96900e+02, 2.36000e+01],\n",
              "        [4.55587e+00, 0.00000e+00, 1.81000e+01, ..., 2.02000e+01,\n",
              "         3.54700e+02, 7.12000e+00]]))"
            ]
          },
          "metadata": {},
          "execution_count": 3
        }
      ],
      "source": [
        "X_train, X_test"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "4WQQZfgfRrgS"
      },
      "outputs": [],
      "source": [
        "scaler = StandardScaler()\n",
        "scaler.fit(X_train)\n",
        "X_train = scaler.transform(X_train)\n",
        "X_test = scaler.transform(X_test)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "5IvCI4ZnSPbv",
        "outputId": "3f22a7f6-674a-40ac-a4c8-d52ab9970029"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(array([[ 5.85233929, -0.48325508,  0.96757729, ...,  0.82585986,\n",
              "         -3.49352799, -0.3830306 ],\n",
              "        [-0.47378382, -0.48325508, -1.01913335, ...,  0.06092409,\n",
              "          0.41561094, -0.88066859],\n",
              "        [-0.47645529,  0.39893917, -1.18846358, ..., -1.5589399 ,\n",
              "          0.3655555 , -1.26571251],\n",
              "        ...,\n",
              "        [ 0.69738702, -0.48325508,  0.96757729, ...,  0.82585986,\n",
              "         -3.4526009 ,  0.81241261],\n",
              "        [-0.17750966, -0.48325508,  1.1836193 , ..., -1.64893234,\n",
              "         -0.20892842, -0.21483452],\n",
              "        [-0.45654127,  0.48715859, -0.81914852, ...,  0.33090142,\n",
              "          0.25350756, -0.3774704 ]]),\n",
              " array([[-0.40514677,  0.39893917, -1.09504001, ..., -2.41386811,\n",
              "          0.38952995, -1.07666568],\n",
              "        [-0.4543714 ,  0.48715859, -0.81914852, ...,  0.33090142,\n",
              "          0.26283653, -1.28934336],\n",
              "        [-0.46965958,  0.39893917, -0.65857676, ...,  0.10592031,\n",
              "          0.40598103,  0.10904719],\n",
              "        ...,\n",
              "        [-0.4373976 , -0.48325508, -0.59726754, ...,  0.55588253,\n",
              "          0.46175422, -0.71803271],\n",
              "        [ 0.65693601, -0.48325508,  0.96757729, ...,  0.82585986,\n",
              "          0.46175422,  1.49214718],\n",
              "        [ 0.0833629 , -0.48325508,  0.96757729, ...,  0.82585986,\n",
              "          0.03843972, -0.79865563]]))"
            ]
          },
          "metadata": {},
          "execution_count": 5
        }
      ],
      "source": [
        "X_train, X_test"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "KUBVrJAl7Z6R"
      },
      "source": [
        "# Kreiranje mreže"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "38uklXzMcxtA"
      },
      "outputs": [],
      "source": [
        "input_size = X_train.shape[1]\n",
        "output_size = 1\n",
        "regularizer = regularizers.l2()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "KKmVN605p96O"
      },
      "outputs": [],
      "source": [
        "model = Sequential([\n",
        "    InputLayer(input_shape=((input_size,))),\n",
        "    Dense(units=NUM_NEURONS_1, activation=\"relu\"),\n",
        "    Dense(units=NUM_NEURONS_2, activation=\"relu\"),\n",
        "    Dense(units=1)\n",
        "])\n",
        "model_reg = Sequential([\n",
        "    InputLayer(input_shape=((input_size,))),\n",
        "    Dense(units=NUM_NEURONS_1, activation=\"relu\", kernel_regularizer=regularizer),\n",
        "    Dense(units=NUM_NEURONS_2, activation=\"relu\", kernel_regularizer=regularizer),\n",
        "    Dense(units=1)\n",
        "])\n",
        "model_drop = Sequential([\n",
        "    InputLayer(input_shape=((input_size,))),\n",
        "    Dense(units=NUM_NEURONS_1, activation=\"relu\"),\n",
        "    Dropout(rate=0.3),\n",
        "    Dense(units=NUM_NEURONS_2, activation=\"relu\"),\n",
        "    Dropout(rate=0.2),\n",
        "    Dense(units=1)\n",
        "])\n",
        "model_reg_drop = Sequential([\n",
        "    InputLayer(input_shape=((input_size,))),\n",
        "    Dense(units=NUM_NEURONS_1, activation=\"relu\", kernel_regularizer=regularizer),\n",
        "    Dropout(rate=0.3),\n",
        "    Dense(units=NUM_NEURONS_2, activation=\"relu\", kernel_regularizer=regularizer),\n",
        "    Dropout(rate=0.2),\n",
        "    Dense(units=1)\n",
        "])"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "oM_1_-2wG84o",
        "outputId": "08d06226-c3b0-462a-94fe-49827683fdf1"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Model: \"sequential\"\n",
            "_________________________________________________________________\n",
            " Layer (type)                Output Shape              Param #   \n",
            "=================================================================\n",
            " dense (Dense)               (None, 200)               2800      \n",
            "                                                                 \n",
            " dense_1 (Dense)             (None, 100)               20100     \n",
            "                                                                 \n",
            " dense_2 (Dense)             (None, 1)                 101       \n",
            "                                                                 \n",
            "=================================================================\n",
            "Total params: 23001 (89.85 KB)\n",
            "Trainable params: 23001 (89.85 KB)\n",
            "Non-trainable params: 0 (0.00 Byte)\n",
            "_________________________________________________________________\n"
          ]
        }
      ],
      "source": [
        "model.summary()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "3X2ExxyMqG72",
        "outputId": "8ab268d9-b284-4f97-c7e4-e5bb731bfdfa"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Model: \"sequential_1\"\n",
            "_________________________________________________________________\n",
            " Layer (type)                Output Shape              Param #   \n",
            "=================================================================\n",
            " dense_3 (Dense)             (None, 200)               2800      \n",
            "                                                                 \n",
            " dense_4 (Dense)             (None, 100)               20100     \n",
            "                                                                 \n",
            " dense_5 (Dense)             (None, 1)                 101       \n",
            "                                                                 \n",
            "=================================================================\n",
            "Total params: 23001 (89.85 KB)\n",
            "Trainable params: 23001 (89.85 KB)\n",
            "Non-trainable params: 0 (0.00 Byte)\n",
            "_________________________________________________________________\n"
          ]
        }
      ],
      "source": [
        "model_reg.summary()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "nl4wXXirtzCH",
        "outputId": "078fb228-0959-4e4f-83a7-997a8d467f01"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Model: \"sequential_2\"\n",
            "_________________________________________________________________\n",
            " Layer (type)                Output Shape              Param #   \n",
            "=================================================================\n",
            " dense_6 (Dense)             (None, 200)               2800      \n",
            "                                                                 \n",
            " dropout (Dropout)           (None, 200)               0         \n",
            "                                                                 \n",
            " dense_7 (Dense)             (None, 100)               20100     \n",
            "                                                                 \n",
            " dropout_1 (Dropout)         (None, 100)               0         \n",
            "                                                                 \n",
            " dense_8 (Dense)             (None, 1)                 101       \n",
            "                                                                 \n",
            "=================================================================\n",
            "Total params: 23001 (89.85 KB)\n",
            "Trainable params: 23001 (89.85 KB)\n",
            "Non-trainable params: 0 (0.00 Byte)\n",
            "_________________________________________________________________\n"
          ]
        }
      ],
      "source": [
        "model_drop.summary()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "0fu5uEMGHWvz",
        "outputId": "28488716-01b4-422c-dd15-90c0e2d712fb"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Model: \"sequential_3\"\n",
            "_________________________________________________________________\n",
            " Layer (type)                Output Shape              Param #   \n",
            "=================================================================\n",
            " dense_9 (Dense)             (None, 200)               2800      \n",
            "                                                                 \n",
            " dropout_2 (Dropout)         (None, 200)               0         \n",
            "                                                                 \n",
            " dense_10 (Dense)            (None, 100)               20100     \n",
            "                                                                 \n",
            " dropout_3 (Dropout)         (None, 100)               0         \n",
            "                                                                 \n",
            " dense_11 (Dense)            (None, 1)                 101       \n",
            "                                                                 \n",
            "=================================================================\n",
            "Total params: 23001 (89.85 KB)\n",
            "Trainable params: 23001 (89.85 KB)\n",
            "Non-trainable params: 0 (0.00 Byte)\n",
            "_________________________________________________________________\n"
          ]
        }
      ],
      "source": [
        "model_reg_drop.summary()"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "pwYNsCS0Tll0"
      },
      "source": [
        "#Treniranje mreže"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "zB3ewXlUHHTe"
      },
      "outputs": [],
      "source": [
        "model.compile(optimizer=Adam(learning_rate=0.001), loss=MeanSquaredError(), metrics=[MeanAbsoluteError()])\n",
        "model_reg.compile(optimizer=Adam(learning_rate=0.001), loss=MeanSquaredError(), metrics=[MeanAbsoluteError()])\n",
        "model_drop.compile(optimizer=Adam(learning_rate=0.001), loss=MeanSquaredError(), metrics=[MeanAbsoluteError()])\n",
        "model_reg_drop.compile(optimizer=Adam(learning_rate=0.001), loss=MeanSquaredError(), metrics=[MeanAbsoluteError()])"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "oBE1c5HoBuJx"
      },
      "outputs": [],
      "source": [
        "early_stop = EarlyStopping(monitor=\"val_loss\", patience=30, verbose=1)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "3ocMfJCyHpAT",
        "outputId": "11629b6e-bb3c-49cd-f12e-0007b10a0987"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/400\n",
            "5/5 [==============================] - 4s 254ms/step - loss: 567.6113 - mean_absolute_error: 21.3117 - val_loss: 579.7986 - val_mean_absolute_error: 20.7564\n",
            "Epoch 2/400\n",
            "5/5 [==============================] - 0s 56ms/step - loss: 525.6254 - mean_absolute_error: 21.2271 - val_loss: 538.1700 - val_mean_absolute_error: 19.7724\n",
            "Epoch 3/400\n",
            "5/5 [==============================] - 0s 68ms/step - loss: 483.0569 - mean_absolute_error: 19.8059 - val_loss: 491.0954 - val_mean_absolute_error: 18.6390\n",
            "Epoch 4/400\n",
            "5/5 [==============================] - 0s 29ms/step - loss: 433.6060 - mean_absolute_error: 18.8116 - val_loss: 436.7868 - val_mean_absolute_error: 17.3248\n",
            "Epoch 5/400\n",
            "5/5 [==============================] - 0s 47ms/step - loss: 377.0715 - mean_absolute_error: 17.3100 - val_loss: 373.2857 - val_mean_absolute_error: 15.7436\n",
            "Epoch 6/400\n",
            "5/5 [==============================] - 0s 75ms/step - loss: 311.5813 - mean_absolute_error: 14.7999 - val_loss: 303.2470 - val_mean_absolute_error: 13.8727\n",
            "Epoch 7/400\n",
            "5/5 [==============================] - 0s 61ms/step - loss: 242.6714 - mean_absolute_error: 13.5902 - val_loss: 231.7848 - val_mean_absolute_error: 11.7294\n",
            "Epoch 8/400\n",
            "5/5 [==============================] - 0s 45ms/step - loss: 174.2856 - mean_absolute_error: 10.5080 - val_loss: 165.5183 - val_mean_absolute_error: 9.2799\n",
            "Epoch 9/400\n",
            "5/5 [==============================] - 0s 23ms/step - loss: 115.1555 - mean_absolute_error: 8.4551 - val_loss: 113.2050 - val_mean_absolute_error: 6.7806\n",
            "Epoch 10/400\n",
            "5/5 [==============================] - 1s 257ms/step - loss: 74.4285 - mean_absolute_error: 6.6959 - val_loss: 81.0484 - val_mean_absolute_error: 4.9358\n",
            "Epoch 11/400\n",
            "5/5 [==============================] - 0s 23ms/step - loss: 55.7963 - mean_absolute_error: 5.5311 - val_loss: 67.0534 - val_mean_absolute_error: 4.3960\n",
            "Epoch 12/400\n",
            "5/5 [==============================] - 0s 68ms/step - loss: 48.1918 - mean_absolute_error: 5.0120 - val_loss: 60.9368 - val_mean_absolute_error: 4.2783\n",
            "Epoch 13/400\n",
            "5/5 [==============================] - 0s 21ms/step - loss: 43.0315 - mean_absolute_error: 4.8345 - val_loss: 54.2775 - val_mean_absolute_error: 3.7752\n",
            "Epoch 14/400\n",
            "5/5 [==============================] - 0s 53ms/step - loss: 35.7758 - mean_absolute_error: 4.4046 - val_loss: 47.7019 - val_mean_absolute_error: 3.0676\n",
            "Epoch 15/400\n",
            "5/5 [==============================] - 0s 49ms/step - loss: 29.0475 - mean_absolute_error: 4.2577 - val_loss: 43.5904 - val_mean_absolute_error: 3.2744\n",
            "Epoch 16/400\n",
            "5/5 [==============================] - 0s 57ms/step - loss: 24.9456 - mean_absolute_error: 3.5722 - val_loss: 41.9585 - val_mean_absolute_error: 3.5121\n",
            "Epoch 17/400\n",
            "5/5 [==============================] - 0s 29ms/step - loss: 23.2400 - mean_absolute_error: 3.6754 - val_loss: 41.1163 - val_mean_absolute_error: 3.6581\n",
            "Epoch 18/400\n",
            "5/5 [==============================] - 0s 25ms/step - loss: 22.0071 - mean_absolute_error: 3.5854 - val_loss: 39.5863 - val_mean_absolute_error: 3.6464\n",
            "Epoch 19/400\n",
            "5/5 [==============================] - 0s 40ms/step - loss: 20.8779 - mean_absolute_error: 3.3173 - val_loss: 38.1963 - val_mean_absolute_error: 3.5491\n",
            "Epoch 20/400\n",
            "5/5 [==============================] - 0s 46ms/step - loss: 19.6079 - mean_absolute_error: 3.2931 - val_loss: 36.3639 - val_mean_absolute_error: 3.4061\n",
            "Epoch 21/400\n",
            "5/5 [==============================] - 0s 67ms/step - loss: 18.7615 - mean_absolute_error: 3.4028 - val_loss: 35.0175 - val_mean_absolute_error: 3.3085\n",
            "Epoch 22/400\n",
            "5/5 [==============================] - 0s 28ms/step - loss: 18.2077 - mean_absolute_error: 3.0331 - val_loss: 34.5568 - val_mean_absolute_error: 3.2865\n",
            "Epoch 23/400\n",
            "5/5 [==============================] - 0s 32ms/step - loss: 17.4199 - mean_absolute_error: 3.0155 - val_loss: 33.7756 - val_mean_absolute_error: 3.3134\n",
            "Epoch 24/400\n",
            "5/5 [==============================] - 0s 36ms/step - loss: 16.8286 - mean_absolute_error: 2.8835 - val_loss: 33.2306 - val_mean_absolute_error: 3.3455\n",
            "Epoch 25/400\n",
            "5/5 [==============================] - 0s 48ms/step - loss: 16.4152 - mean_absolute_error: 2.7693 - val_loss: 32.7425 - val_mean_absolute_error: 3.3338\n",
            "Epoch 26/400\n",
            "5/5 [==============================] - 0s 72ms/step - loss: 15.9833 - mean_absolute_error: 2.8774 - val_loss: 32.1354 - val_mean_absolute_error: 3.2957\n",
            "Epoch 27/400\n",
            "5/5 [==============================] - 0s 33ms/step - loss: 15.7181 - mean_absolute_error: 2.9443 - val_loss: 31.8170 - val_mean_absolute_error: 3.2950\n",
            "Epoch 28/400\n",
            "5/5 [==============================] - 0s 34ms/step - loss: 15.3344 - mean_absolute_error: 2.9856 - val_loss: 31.0854 - val_mean_absolute_error: 3.2272\n",
            "Epoch 29/400\n",
            "5/5 [==============================] - 0s 27ms/step - loss: 14.9296 - mean_absolute_error: 2.7800 - val_loss: 30.0961 - val_mean_absolute_error: 3.1547\n",
            "Epoch 30/400\n",
            "5/5 [==============================] - 0s 28ms/step - loss: 14.6078 - mean_absolute_error: 2.6850 - val_loss: 29.3402 - val_mean_absolute_error: 3.0998\n",
            "Epoch 31/400\n",
            "5/5 [==============================] - 0s 47ms/step - loss: 14.2699 - mean_absolute_error: 2.6536 - val_loss: 28.4799 - val_mean_absolute_error: 3.0895\n",
            "Epoch 32/400\n",
            "5/5 [==============================] - 0s 73ms/step - loss: 13.9887 - mean_absolute_error: 2.7101 - val_loss: 27.8089 - val_mean_absolute_error: 3.0972\n",
            "Epoch 33/400\n",
            "5/5 [==============================] - 0s 46ms/step - loss: 13.6875 - mean_absolute_error: 2.5751 - val_loss: 27.7260 - val_mean_absolute_error: 3.1017\n",
            "Epoch 34/400\n",
            "5/5 [==============================] - 0s 59ms/step - loss: 13.3899 - mean_absolute_error: 2.4997 - val_loss: 27.5611 - val_mean_absolute_error: 3.1453\n",
            "Epoch 35/400\n",
            "5/5 [==============================] - 0s 99ms/step - loss: 13.2227 - mean_absolute_error: 2.6002 - val_loss: 27.2075 - val_mean_absolute_error: 3.1455\n",
            "Epoch 36/400\n",
            "5/5 [==============================] - 0s 51ms/step - loss: 12.9320 - mean_absolute_error: 2.4893 - val_loss: 26.6549 - val_mean_absolute_error: 3.0956\n",
            "Epoch 37/400\n",
            "5/5 [==============================] - 0s 63ms/step - loss: 12.7029 - mean_absolute_error: 2.4394 - val_loss: 26.1464 - val_mean_absolute_error: 3.0689\n",
            "Epoch 38/400\n",
            "5/5 [==============================] - 0s 77ms/step - loss: 12.5111 - mean_absolute_error: 2.6318 - val_loss: 25.5885 - val_mean_absolute_error: 3.0216\n",
            "Epoch 39/400\n",
            "5/5 [==============================] - 0s 45ms/step - loss: 12.3686 - mean_absolute_error: 2.5560 - val_loss: 25.4489 - val_mean_absolute_error: 2.9797\n",
            "Epoch 40/400\n",
            "5/5 [==============================] - 0s 48ms/step - loss: 12.2033 - mean_absolute_error: 2.5409 - val_loss: 25.0790 - val_mean_absolute_error: 3.0157\n",
            "Epoch 41/400\n",
            "5/5 [==============================] - 0s 70ms/step - loss: 12.0045 - mean_absolute_error: 2.3971 - val_loss: 25.1644 - val_mean_absolute_error: 3.0775\n",
            "Epoch 42/400\n",
            "5/5 [==============================] - 0s 48ms/step - loss: 11.8013 - mean_absolute_error: 2.5164 - val_loss: 24.9383 - val_mean_absolute_error: 3.1023\n",
            "Epoch 43/400\n",
            "5/5 [==============================] - 0s 60ms/step - loss: 11.5851 - mean_absolute_error: 2.3999 - val_loss: 24.6108 - val_mean_absolute_error: 3.1147\n",
            "Epoch 44/400\n",
            "5/5 [==============================] - 0s 50ms/step - loss: 11.4326 - mean_absolute_error: 2.4697 - val_loss: 24.4449 - val_mean_absolute_error: 3.1123\n",
            "Epoch 45/400\n",
            "5/5 [==============================] - 0s 66ms/step - loss: 11.3868 - mean_absolute_error: 2.6571 - val_loss: 24.0959 - val_mean_absolute_error: 3.0647\n",
            "Epoch 46/400\n",
            "5/5 [==============================] - 0s 46ms/step - loss: 11.1440 - mean_absolute_error: 2.3544 - val_loss: 23.4759 - val_mean_absolute_error: 2.9938\n",
            "Epoch 47/400\n",
            "5/5 [==============================] - 0s 47ms/step - loss: 10.9921 - mean_absolute_error: 2.3480 - val_loss: 23.1309 - val_mean_absolute_error: 2.9545\n",
            "Epoch 48/400\n",
            "5/5 [==============================] - 0s 61ms/step - loss: 10.9170 - mean_absolute_error: 2.4136 - val_loss: 22.8550 - val_mean_absolute_error: 2.9297\n",
            "Epoch 49/400\n",
            "5/5 [==============================] - 0s 38ms/step - loss: 10.7681 - mean_absolute_error: 2.3553 - val_loss: 22.6477 - val_mean_absolute_error: 2.9254\n",
            "Epoch 50/400\n",
            "5/5 [==============================] - 0s 50ms/step - loss: 10.6607 - mean_absolute_error: 2.3720 - val_loss: 22.5464 - val_mean_absolute_error: 2.9387\n",
            "Epoch 51/400\n",
            "5/5 [==============================] - 0s 68ms/step - loss: 10.5093 - mean_absolute_error: 2.2921 - val_loss: 22.5142 - val_mean_absolute_error: 2.9375\n",
            "Epoch 52/400\n",
            "5/5 [==============================] - 0s 40ms/step - loss: 10.4203 - mean_absolute_error: 2.4499 - val_loss: 22.2891 - val_mean_absolute_error: 2.9137\n",
            "Epoch 53/400\n",
            "5/5 [==============================] - 0s 23ms/step - loss: 10.2667 - mean_absolute_error: 2.3578 - val_loss: 21.9563 - val_mean_absolute_error: 2.9270\n",
            "Epoch 54/400\n",
            "5/5 [==============================] - 0s 50ms/step - loss: 10.2765 - mean_absolute_error: 2.3312 - val_loss: 21.8078 - val_mean_absolute_error: 2.9315\n",
            "Epoch 55/400\n",
            "5/5 [==============================] - 0s 88ms/step - loss: 10.1548 - mean_absolute_error: 2.3999 - val_loss: 21.9058 - val_mean_absolute_error: 2.9690\n",
            "Epoch 56/400\n",
            "5/5 [==============================] - 0s 69ms/step - loss: 9.9549 - mean_absolute_error: 2.3024 - val_loss: 22.0365 - val_mean_absolute_error: 2.9917\n",
            "Epoch 57/400\n",
            "5/5 [==============================] - 0s 42ms/step - loss: 9.8667 - mean_absolute_error: 2.2637 - val_loss: 22.2137 - val_mean_absolute_error: 3.0347\n",
            "Epoch 58/400\n",
            "5/5 [==============================] - 0s 43ms/step - loss: 9.8013 - mean_absolute_error: 2.2601 - val_loss: 21.9759 - val_mean_absolute_error: 3.0478\n",
            "Epoch 59/400\n",
            "5/5 [==============================] - 0s 15ms/step - loss: 9.7369 - mean_absolute_error: 2.3128 - val_loss: 21.7077 - val_mean_absolute_error: 3.0586\n",
            "Epoch 60/400\n",
            "5/5 [==============================] - 0s 15ms/step - loss: 9.6614 - mean_absolute_error: 2.2119 - val_loss: 21.4800 - val_mean_absolute_error: 3.0732\n",
            "Epoch 61/400\n",
            "5/5 [==============================] - 0s 11ms/step - loss: 9.5597 - mean_absolute_error: 2.2528 - val_loss: 21.3610 - val_mean_absolute_error: 3.0777\n",
            "Epoch 62/400\n",
            "5/5 [==============================] - 0s 12ms/step - loss: 9.4575 - mean_absolute_error: 2.2450 - val_loss: 21.3939 - val_mean_absolute_error: 3.0325\n",
            "Epoch 63/400\n",
            "5/5 [==============================] - 0s 17ms/step - loss: 9.3976 - mean_absolute_error: 2.1675 - val_loss: 21.3551 - val_mean_absolute_error: 2.9869\n",
            "Epoch 64/400\n",
            "5/5 [==============================] - 0s 16ms/step - loss: 9.3410 - mean_absolute_error: 2.1835 - val_loss: 21.7053 - val_mean_absolute_error: 2.9762\n",
            "Epoch 65/400\n",
            "5/5 [==============================] - 0s 12ms/step - loss: 9.2905 - mean_absolute_error: 2.1341 - val_loss: 21.5128 - val_mean_absolute_error: 2.9213\n",
            "Epoch 66/400\n",
            "5/5 [==============================] - 0s 12ms/step - loss: 9.2088 - mean_absolute_error: 2.1906 - val_loss: 21.0882 - val_mean_absolute_error: 2.8565\n",
            "Epoch 67/400\n",
            "5/5 [==============================] - 0s 17ms/step - loss: 9.1180 - mean_absolute_error: 2.3009 - val_loss: 20.7283 - val_mean_absolute_error: 2.8073\n",
            "Epoch 68/400\n",
            "5/5 [==============================] - 0s 14ms/step - loss: 9.0434 - mean_absolute_error: 2.1237 - val_loss: 20.6435 - val_mean_absolute_error: 2.8364\n",
            "Epoch 69/400\n",
            "5/5 [==============================] - 0s 14ms/step - loss: 8.9222 - mean_absolute_error: 2.1414 - val_loss: 20.8406 - val_mean_absolute_error: 2.8719\n",
            "Epoch 70/400\n",
            "5/5 [==============================] - 0s 12ms/step - loss: 8.9186 - mean_absolute_error: 2.1813 - val_loss: 21.2095 - val_mean_absolute_error: 2.9407\n",
            "Epoch 71/400\n",
            "5/5 [==============================] - 0s 12ms/step - loss: 8.9041 - mean_absolute_error: 2.3006 - val_loss: 20.8080 - val_mean_absolute_error: 2.9149\n",
            "Epoch 72/400\n",
            "5/5 [==============================] - 0s 16ms/step - loss: 8.7106 - mean_absolute_error: 2.1387 - val_loss: 20.3017 - val_mean_absolute_error: 2.8578\n",
            "Epoch 73/400\n",
            "5/5 [==============================] - 0s 12ms/step - loss: 8.7325 - mean_absolute_error: 2.2529 - val_loss: 20.1365 - val_mean_absolute_error: 2.8577\n",
            "Epoch 74/400\n",
            "5/5 [==============================] - 0s 13ms/step - loss: 8.7358 - mean_absolute_error: 2.1483 - val_loss: 20.1595 - val_mean_absolute_error: 2.8814\n",
            "Epoch 75/400\n",
            "5/5 [==============================] - 0s 12ms/step - loss: 8.5632 - mean_absolute_error: 2.0782 - val_loss: 20.3388 - val_mean_absolute_error: 2.9243\n",
            "Epoch 76/400\n",
            "5/5 [==============================] - 0s 12ms/step - loss: 8.4797 - mean_absolute_error: 2.0925 - val_loss: 20.3994 - val_mean_absolute_error: 2.9699\n",
            "Epoch 77/400\n",
            "5/5 [==============================] - 0s 16ms/step - loss: 8.4165 - mean_absolute_error: 2.0500 - val_loss: 20.3336 - val_mean_absolute_error: 2.9659\n",
            "Epoch 78/400\n",
            "5/5 [==============================] - 0s 16ms/step - loss: 8.3690 - mean_absolute_error: 2.1513 - val_loss: 20.1951 - val_mean_absolute_error: 2.9254\n",
            "Epoch 79/400\n",
            "5/5 [==============================] - 0s 12ms/step - loss: 8.2959 - mean_absolute_error: 2.0422 - val_loss: 20.1305 - val_mean_absolute_error: 2.8774\n",
            "Epoch 80/400\n",
            "5/5 [==============================] - 0s 15ms/step - loss: 8.1735 - mean_absolute_error: 2.1519 - val_loss: 19.9501 - val_mean_absolute_error: 2.8512\n",
            "Epoch 81/400\n",
            "5/5 [==============================] - 0s 16ms/step - loss: 8.1300 - mean_absolute_error: 2.1991 - val_loss: 19.8586 - val_mean_absolute_error: 2.8244\n",
            "Epoch 82/400\n",
            "5/5 [==============================] - 0s 11ms/step - loss: 8.2899 - mean_absolute_error: 2.1747 - val_loss: 20.4141 - val_mean_absolute_error: 2.8650\n",
            "Epoch 83/400\n",
            "5/5 [==============================] - 0s 19ms/step - loss: 8.1833 - mean_absolute_error: 2.0698 - val_loss: 20.3330 - val_mean_absolute_error: 2.9202\n",
            "Epoch 84/400\n",
            "5/5 [==============================] - 0s 15ms/step - loss: 7.9706 - mean_absolute_error: 2.0585 - val_loss: 20.0926 - val_mean_absolute_error: 2.9469\n",
            "Epoch 85/400\n",
            "5/5 [==============================] - 0s 11ms/step - loss: 7.9092 - mean_absolute_error: 2.0046 - val_loss: 19.9405 - val_mean_absolute_error: 2.9883\n",
            "Epoch 86/400\n",
            "5/5 [==============================] - 0s 11ms/step - loss: 7.8853 - mean_absolute_error: 2.0916 - val_loss: 19.7929 - val_mean_absolute_error: 2.9885\n",
            "Epoch 87/400\n",
            "5/5 [==============================] - 0s 11ms/step - loss: 7.8436 - mean_absolute_error: 2.1099 - val_loss: 19.4962 - val_mean_absolute_error: 2.9202\n",
            "Epoch 88/400\n",
            "5/5 [==============================] - 0s 16ms/step - loss: 7.7561 - mean_absolute_error: 2.0019 - val_loss: 19.4722 - val_mean_absolute_error: 2.9089\n",
            "Epoch 89/400\n",
            "5/5 [==============================] - 0s 12ms/step - loss: 7.6747 - mean_absolute_error: 2.0037 - val_loss: 19.5130 - val_mean_absolute_error: 2.8986\n",
            "Epoch 90/400\n",
            "5/5 [==============================] - 0s 11ms/step - loss: 7.6224 - mean_absolute_error: 2.0471 - val_loss: 19.5256 - val_mean_absolute_error: 2.9159\n",
            "Epoch 91/400\n",
            "5/5 [==============================] - 0s 11ms/step - loss: 7.5537 - mean_absolute_error: 1.9713 - val_loss: 19.3899 - val_mean_absolute_error: 2.9027\n",
            "Epoch 92/400\n",
            "5/5 [==============================] - 0s 11ms/step - loss: 7.5135 - mean_absolute_error: 2.0835 - val_loss: 19.2503 - val_mean_absolute_error: 2.8633\n",
            "Epoch 93/400\n",
            "5/5 [==============================] - 0s 12ms/step - loss: 7.5679 - mean_absolute_error: 1.9799 - val_loss: 19.0381 - val_mean_absolute_error: 2.7934\n",
            "Epoch 94/400\n",
            "5/5 [==============================] - 0s 12ms/step - loss: 7.5122 - mean_absolute_error: 2.0843 - val_loss: 19.5203 - val_mean_absolute_error: 2.8189\n",
            "Epoch 95/400\n",
            "5/5 [==============================] - 0s 12ms/step - loss: 7.3734 - mean_absolute_error: 1.9933 - val_loss: 20.0542 - val_mean_absolute_error: 2.9337\n",
            "Epoch 96/400\n",
            "5/5 [==============================] - 0s 11ms/step - loss: 7.4173 - mean_absolute_error: 1.9594 - val_loss: 20.3887 - val_mean_absolute_error: 3.0087\n",
            "Epoch 97/400\n",
            "5/5 [==============================] - 0s 14ms/step - loss: 7.4232 - mean_absolute_error: 2.0331 - val_loss: 20.0588 - val_mean_absolute_error: 2.9371\n",
            "Epoch 98/400\n",
            "5/5 [==============================] - 0s 16ms/step - loss: 7.3715 - mean_absolute_error: 1.9754 - val_loss: 19.4424 - val_mean_absolute_error: 2.8316\n",
            "Epoch 99/400\n",
            "5/5 [==============================] - 0s 14ms/step - loss: 7.2625 - mean_absolute_error: 2.0437 - val_loss: 19.6066 - val_mean_absolute_error: 2.8018\n",
            "Epoch 100/400\n",
            "5/5 [==============================] - 0s 11ms/step - loss: 7.1666 - mean_absolute_error: 1.9523 - val_loss: 19.8824 - val_mean_absolute_error: 2.8180\n",
            "Epoch 101/400\n",
            "5/5 [==============================] - 0s 12ms/step - loss: 7.1722 - mean_absolute_error: 2.0376 - val_loss: 20.1293 - val_mean_absolute_error: 2.8793\n",
            "Epoch 102/400\n",
            "5/5 [==============================] - 0s 15ms/step - loss: 7.1386 - mean_absolute_error: 1.9549 - val_loss: 19.8953 - val_mean_absolute_error: 2.8951\n",
            "Epoch 103/400\n",
            "5/5 [==============================] - 0s 12ms/step - loss: 7.0270 - mean_absolute_error: 1.9078 - val_loss: 19.3147 - val_mean_absolute_error: 2.8576\n",
            "Epoch 104/400\n",
            "5/5 [==============================] - 0s 12ms/step - loss: 7.0187 - mean_absolute_error: 1.9292 - val_loss: 19.0223 - val_mean_absolute_error: 2.7847\n",
            "Epoch 105/400\n",
            "5/5 [==============================] - 0s 11ms/step - loss: 6.9887 - mean_absolute_error: 1.9303 - val_loss: 19.1335 - val_mean_absolute_error: 2.7665\n",
            "Epoch 106/400\n",
            "5/5 [==============================] - 0s 16ms/step - loss: 6.9518 - mean_absolute_error: 2.0680 - val_loss: 19.1949 - val_mean_absolute_error: 2.7752\n",
            "Epoch 107/400\n",
            "5/5 [==============================] - 0s 16ms/step - loss: 6.8349 - mean_absolute_error: 2.0257 - val_loss: 18.7108 - val_mean_absolute_error: 2.7622\n",
            "Epoch 108/400\n",
            "5/5 [==============================] - 0s 16ms/step - loss: 6.8736 - mean_absolute_error: 2.0350 - val_loss: 18.2926 - val_mean_absolute_error: 2.6925\n",
            "Epoch 109/400\n",
            "5/5 [==============================] - 0s 15ms/step - loss: 6.8740 - mean_absolute_error: 2.0581 - val_loss: 18.4717 - val_mean_absolute_error: 2.6804\n",
            "Epoch 110/400\n",
            "5/5 [==============================] - 0s 12ms/step - loss: 6.7653 - mean_absolute_error: 1.8222 - val_loss: 18.6594 - val_mean_absolute_error: 2.7494\n",
            "Epoch 111/400\n",
            "5/5 [==============================] - 0s 12ms/step - loss: 6.6916 - mean_absolute_error: 1.9437 - val_loss: 18.9057 - val_mean_absolute_error: 2.8177\n",
            "Epoch 112/400\n",
            "5/5 [==============================] - 0s 17ms/step - loss: 6.6723 - mean_absolute_error: 1.8699 - val_loss: 19.2251 - val_mean_absolute_error: 2.8029\n",
            "Epoch 113/400\n",
            "5/5 [==============================] - 0s 16ms/step - loss: 6.6138 - mean_absolute_error: 2.0858 - val_loss: 19.0270 - val_mean_absolute_error: 2.7844\n",
            "Epoch 114/400\n",
            "5/5 [==============================] - 0s 18ms/step - loss: 6.5393 - mean_absolute_error: 1.9459 - val_loss: 18.5357 - val_mean_absolute_error: 2.7373\n",
            "Epoch 115/400\n",
            "5/5 [==============================] - 0s 15ms/step - loss: 6.4946 - mean_absolute_error: 1.8571 - val_loss: 18.4714 - val_mean_absolute_error: 2.7287\n",
            "Epoch 116/400\n",
            "5/5 [==============================] - 0s 16ms/step - loss: 6.5113 - mean_absolute_error: 1.9220 - val_loss: 18.6285 - val_mean_absolute_error: 2.7124\n",
            "Epoch 117/400\n",
            "5/5 [==============================] - 0s 16ms/step - loss: 6.4513 - mean_absolute_error: 2.0272 - val_loss: 18.9038 - val_mean_absolute_error: 2.7042\n",
            "Epoch 118/400\n",
            "5/5 [==============================] - 0s 12ms/step - loss: 6.4172 - mean_absolute_error: 1.9286 - val_loss: 18.8342 - val_mean_absolute_error: 2.7067\n",
            "Epoch 119/400\n",
            "5/5 [==============================] - 0s 13ms/step - loss: 6.2923 - mean_absolute_error: 1.9036 - val_loss: 19.2270 - val_mean_absolute_error: 2.7427\n",
            "Epoch 120/400\n",
            "5/5 [==============================] - 0s 13ms/step - loss: 6.2655 - mean_absolute_error: 1.9010 - val_loss: 19.8412 - val_mean_absolute_error: 2.7813\n",
            "Epoch 121/400\n",
            "5/5 [==============================] - 0s 12ms/step - loss: 6.2836 - mean_absolute_error: 2.1091 - val_loss: 19.8646 - val_mean_absolute_error: 2.7841\n",
            "Epoch 122/400\n",
            "5/5 [==============================] - 0s 11ms/step - loss: 6.2052 - mean_absolute_error: 1.7972 - val_loss: 18.9891 - val_mean_absolute_error: 2.7657\n",
            "Epoch 123/400\n",
            "5/5 [==============================] - 0s 15ms/step - loss: 6.2283 - mean_absolute_error: 1.8940 - val_loss: 18.6723 - val_mean_absolute_error: 2.7662\n",
            "Epoch 124/400\n",
            "5/5 [==============================] - 0s 12ms/step - loss: 6.1069 - mean_absolute_error: 1.9119 - val_loss: 18.8156 - val_mean_absolute_error: 2.7616\n",
            "Epoch 125/400\n",
            "5/5 [==============================] - 0s 16ms/step - loss: 6.0289 - mean_absolute_error: 2.0933 - val_loss: 18.7084 - val_mean_absolute_error: 2.7388\n",
            "Epoch 126/400\n",
            "5/5 [==============================] - 0s 17ms/step - loss: 6.0743 - mean_absolute_error: 1.8737 - val_loss: 18.0600 - val_mean_absolute_error: 2.6612\n",
            "Epoch 127/400\n",
            "5/5 [==============================] - 0s 12ms/step - loss: 6.1416 - mean_absolute_error: 1.8085 - val_loss: 18.1231 - val_mean_absolute_error: 2.7239\n",
            "Epoch 128/400\n",
            "5/5 [==============================] - 0s 11ms/step - loss: 5.9412 - mean_absolute_error: 1.8934 - val_loss: 18.6090 - val_mean_absolute_error: 2.8257\n",
            "Epoch 129/400\n",
            "5/5 [==============================] - 0s 19ms/step - loss: 5.8922 - mean_absolute_error: 1.9085 - val_loss: 18.8606 - val_mean_absolute_error: 2.8613\n",
            "Epoch 130/400\n",
            "5/5 [==============================] - 0s 11ms/step - loss: 5.7829 - mean_absolute_error: 1.8119 - val_loss: 18.6073 - val_mean_absolute_error: 2.7632\n",
            "Epoch 131/400\n",
            "5/5 [==============================] - 0s 11ms/step - loss: 5.7753 - mean_absolute_error: 1.8191 - val_loss: 18.4670 - val_mean_absolute_error: 2.6421\n",
            "Epoch 132/400\n",
            "5/5 [==============================] - 0s 15ms/step - loss: 5.7607 - mean_absolute_error: 1.8029 - val_loss: 18.8191 - val_mean_absolute_error: 2.6780\n",
            "Epoch 133/400\n",
            "5/5 [==============================] - 0s 11ms/step - loss: 5.6495 - mean_absolute_error: 1.8345 - val_loss: 19.2382 - val_mean_absolute_error: 2.7237\n",
            "Epoch 134/400\n",
            "5/5 [==============================] - 0s 16ms/step - loss: 5.7189 - mean_absolute_error: 1.8524 - val_loss: 19.0162 - val_mean_absolute_error: 2.7007\n",
            "Epoch 135/400\n",
            "5/5 [==============================] - 0s 16ms/step - loss: 5.5311 - mean_absolute_error: 1.7752 - val_loss: 18.6092 - val_mean_absolute_error: 2.6036\n",
            "Epoch 136/400\n",
            "5/5 [==============================] - 0s 11ms/step - loss: 5.6069 - mean_absolute_error: 2.0001 - val_loss: 18.4760 - val_mean_absolute_error: 2.5893\n",
            "Epoch 137/400\n",
            "5/5 [==============================] - 0s 11ms/step - loss: 5.4672 - mean_absolute_error: 1.8586 - val_loss: 18.2680 - val_mean_absolute_error: 2.6649\n",
            "Epoch 138/400\n",
            "5/5 [==============================] - 0s 15ms/step - loss: 5.5395 - mean_absolute_error: 1.8332 - val_loss: 18.1832 - val_mean_absolute_error: 2.7183\n",
            "Epoch 139/400\n",
            "5/5 [==============================] - 0s 13ms/step - loss: 5.5557 - mean_absolute_error: 1.7272 - val_loss: 18.1101 - val_mean_absolute_error: 2.6227\n",
            "Epoch 140/400\n",
            "5/5 [==============================] - 0s 11ms/step - loss: 5.4495 - mean_absolute_error: 1.8449 - val_loss: 18.3664 - val_mean_absolute_error: 2.6128\n",
            "Epoch 141/400\n",
            "5/5 [==============================] - 0s 11ms/step - loss: 5.4103 - mean_absolute_error: 1.8453 - val_loss: 18.5792 - val_mean_absolute_error: 2.6325\n",
            "Epoch 142/400\n",
            "5/5 [==============================] - 0s 12ms/step - loss: 5.3209 - mean_absolute_error: 1.7377 - val_loss: 18.8024 - val_mean_absolute_error: 2.6290\n",
            "Epoch 143/400\n",
            "5/5 [==============================] - 0s 18ms/step - loss: 5.2374 - mean_absolute_error: 1.9423 - val_loss: 19.1478 - val_mean_absolute_error: 2.6678\n",
            "Epoch 144/400\n",
            "5/5 [==============================] - 0s 16ms/step - loss: 5.2097 - mean_absolute_error: 1.7756 - val_loss: 19.2829 - val_mean_absolute_error: 2.6940\n",
            "Epoch 145/400\n",
            "5/5 [==============================] - 0s 18ms/step - loss: 5.2194 - mean_absolute_error: 1.7470 - val_loss: 19.1407 - val_mean_absolute_error: 2.6600\n",
            "Epoch 146/400\n",
            "5/5 [==============================] - 0s 16ms/step - loss: 5.1029 - mean_absolute_error: 1.7438 - val_loss: 19.1300 - val_mean_absolute_error: 2.6168\n",
            "Epoch 147/400\n",
            "5/5 [==============================] - 0s 16ms/step - loss: 5.0576 - mean_absolute_error: 1.6555 - val_loss: 19.3099 - val_mean_absolute_error: 2.6013\n",
            "Epoch 148/400\n",
            "5/5 [==============================] - 0s 15ms/step - loss: 5.0252 - mean_absolute_error: 1.7008 - val_loss: 19.2731 - val_mean_absolute_error: 2.6122\n",
            "Epoch 149/400\n",
            "5/5 [==============================] - 0s 12ms/step - loss: 4.9822 - mean_absolute_error: 1.6612 - val_loss: 19.2433 - val_mean_absolute_error: 2.6487\n",
            "Epoch 150/400\n",
            "5/5 [==============================] - 0s 12ms/step - loss: 4.9553 - mean_absolute_error: 1.6796 - val_loss: 19.0725 - val_mean_absolute_error: 2.6235\n",
            "Epoch 151/400\n",
            "5/5 [==============================] - 0s 16ms/step - loss: 4.9531 - mean_absolute_error: 1.7013 - val_loss: 18.8592 - val_mean_absolute_error: 2.5991\n",
            "Epoch 152/400\n",
            "5/5 [==============================] - 0s 12ms/step - loss: 4.8751 - mean_absolute_error: 1.6640 - val_loss: 19.1343 - val_mean_absolute_error: 2.6069\n",
            "Epoch 153/400\n",
            "5/5 [==============================] - 0s 16ms/step - loss: 4.8668 - mean_absolute_error: 1.7246 - val_loss: 19.1913 - val_mean_absolute_error: 2.5932\n",
            "Epoch 154/400\n",
            "5/5 [==============================] - 0s 11ms/step - loss: 4.8604 - mean_absolute_error: 1.7127 - val_loss: 18.9479 - val_mean_absolute_error: 2.5425\n",
            "Epoch 155/400\n",
            "5/5 [==============================] - 0s 11ms/step - loss: 4.8082 - mean_absolute_error: 1.6364 - val_loss: 19.1540 - val_mean_absolute_error: 2.5486\n",
            "Epoch 156/400\n",
            "5/5 [==============================] - 0s 12ms/step - loss: 4.8169 - mean_absolute_error: 1.7958 - val_loss: 19.2432 - val_mean_absolute_error: 2.5694\n",
            "Epoch 156: early stopping\n"
          ]
        }
      ],
      "source": [
        "history = model.fit(X_train, y_train, validation_split=0.2, batch_size=BATCH_SIZE, epochs=NUM_EPOCHS, verbose=1, callbacks=early_stop)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "tA090Ir9qY5-",
        "outputId": "b9b72a8c-3b45-41f9-c5a5-24ede37a87a1"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/400\n",
            "5/5 [==============================] - 1s 60ms/step - loss: 572.5469 - mean_absolute_error: 21.7729 - val_loss: 576.9374 - val_mean_absolute_error: 20.6788\n",
            "Epoch 2/400\n",
            "5/5 [==============================] - 0s 15ms/step - loss: 522.5300 - mean_absolute_error: 21.2463 - val_loss: 527.3881 - val_mean_absolute_error: 19.5234\n",
            "Epoch 3/400\n",
            "5/5 [==============================] - 0s 17ms/step - loss: 473.8583 - mean_absolute_error: 20.1991 - val_loss: 475.2206 - val_mean_absolute_error: 18.3003\n",
            "Epoch 4/400\n",
            "5/5 [==============================] - 0s 15ms/step - loss: 419.9884 - mean_absolute_error: 18.1876 - val_loss: 416.2079 - val_mean_absolute_error: 16.9058\n",
            "Epoch 5/400\n",
            "5/5 [==============================] - 0s 16ms/step - loss: 360.0603 - mean_absolute_error: 16.6154 - val_loss: 350.4976 - val_mean_absolute_error: 15.2464\n",
            "Epoch 6/400\n",
            "5/5 [==============================] - 0s 12ms/step - loss: 292.4305 - mean_absolute_error: 14.1606 - val_loss: 280.8122 - val_mean_absolute_error: 13.3130\n",
            "Epoch 7/400\n",
            "5/5 [==============================] - 0s 11ms/step - loss: 222.2491 - mean_absolute_error: 12.4285 - val_loss: 213.4442 - val_mean_absolute_error: 11.0737\n",
            "Epoch 8/400\n",
            "5/5 [==============================] - 0s 16ms/step - loss: 159.1609 - mean_absolute_error: 9.9230 - val_loss: 156.2713 - val_mean_absolute_error: 8.5985\n",
            "Epoch 9/400\n",
            "5/5 [==============================] - 0s 12ms/step - loss: 109.8367 - mean_absolute_error: 8.2797 - val_loss: 118.5995 - val_mean_absolute_error: 6.5068\n",
            "Epoch 10/400\n",
            "5/5 [==============================] - 0s 16ms/step - loss: 79.5647 - mean_absolute_error: 7.0663 - val_loss: 100.6081 - val_mean_absolute_error: 5.3289\n",
            "Epoch 11/400\n",
            "5/5 [==============================] - 0s 12ms/step - loss: 64.9792 - mean_absolute_error: 6.3121 - val_loss: 91.2498 - val_mean_absolute_error: 5.2643\n",
            "Epoch 12/400\n",
            "5/5 [==============================] - 0s 12ms/step - loss: 57.2237 - mean_absolute_error: 5.6796 - val_loss: 81.1672 - val_mean_absolute_error: 4.7932\n",
            "Epoch 13/400\n",
            "5/5 [==============================] - 0s 20ms/step - loss: 46.6242 - mean_absolute_error: 5.0214 - val_loss: 69.0289 - val_mean_absolute_error: 3.9775\n",
            "Epoch 14/400\n",
            "5/5 [==============================] - 0s 16ms/step - loss: 37.4374 - mean_absolute_error: 4.5776 - val_loss: 59.4354 - val_mean_absolute_error: 3.5934\n",
            "Epoch 15/400\n",
            "5/5 [==============================] - 0s 25ms/step - loss: 31.2506 - mean_absolute_error: 4.3027 - val_loss: 53.0275 - val_mean_absolute_error: 3.6219\n",
            "Epoch 16/400\n",
            "5/5 [==============================] - 0s 15ms/step - loss: 27.1583 - mean_absolute_error: 3.6323 - val_loss: 49.3620 - val_mean_absolute_error: 3.6155\n",
            "Epoch 17/400\n",
            "5/5 [==============================] - 0s 24ms/step - loss: 24.6063 - mean_absolute_error: 3.6128 - val_loss: 46.1721 - val_mean_absolute_error: 3.5779\n",
            "Epoch 18/400\n",
            "5/5 [==============================] - 0s 15ms/step - loss: 22.9668 - mean_absolute_error: 3.4974 - val_loss: 43.8684 - val_mean_absolute_error: 3.5267\n",
            "Epoch 19/400\n",
            "5/5 [==============================] - 0s 18ms/step - loss: 21.6555 - mean_absolute_error: 3.5386 - val_loss: 41.5585 - val_mean_absolute_error: 3.4352\n",
            "Epoch 20/400\n",
            "5/5 [==============================] - 0s 15ms/step - loss: 20.5754 - mean_absolute_error: 3.1578 - val_loss: 40.5200 - val_mean_absolute_error: 3.3665\n",
            "Epoch 21/400\n",
            "5/5 [==============================] - 0s 14ms/step - loss: 19.9400 - mean_absolute_error: 3.2267 - val_loss: 39.6076 - val_mean_absolute_error: 3.3620\n",
            "Epoch 22/400\n",
            "5/5 [==============================] - 0s 19ms/step - loss: 19.2636 - mean_absolute_error: 3.2846 - val_loss: 38.3290 - val_mean_absolute_error: 3.3679\n",
            "Epoch 23/400\n",
            "5/5 [==============================] - 0s 16ms/step - loss: 18.5973 - mean_absolute_error: 3.0046 - val_loss: 37.6715 - val_mean_absolute_error: 3.3656\n",
            "Epoch 24/400\n",
            "5/5 [==============================] - 0s 19ms/step - loss: 18.1158 - mean_absolute_error: 2.8921 - val_loss: 36.7463 - val_mean_absolute_error: 3.3282\n",
            "Epoch 25/400\n",
            "5/5 [==============================] - 0s 19ms/step - loss: 17.6114 - mean_absolute_error: 2.8913 - val_loss: 35.7215 - val_mean_absolute_error: 3.3038\n",
            "Epoch 26/400\n",
            "5/5 [==============================] - 0s 19ms/step - loss: 17.2340 - mean_absolute_error: 2.8933 - val_loss: 34.4033 - val_mean_absolute_error: 3.2557\n",
            "Epoch 27/400\n",
            "5/5 [==============================] - 0s 20ms/step - loss: 16.8138 - mean_absolute_error: 2.8449 - val_loss: 33.5885 - val_mean_absolute_error: 3.2085\n",
            "Epoch 28/400\n",
            "5/5 [==============================] - 0s 19ms/step - loss: 16.4964 - mean_absolute_error: 2.8290 - val_loss: 32.8352 - val_mean_absolute_error: 3.1543\n",
            "Epoch 29/400\n",
            "5/5 [==============================] - 0s 19ms/step - loss: 16.1476 - mean_absolute_error: 2.7144 - val_loss: 32.4344 - val_mean_absolute_error: 3.1345\n",
            "Epoch 30/400\n",
            "5/5 [==============================] - 0s 21ms/step - loss: 15.8428 - mean_absolute_error: 2.6121 - val_loss: 32.1433 - val_mean_absolute_error: 3.1110\n",
            "Epoch 31/400\n",
            "5/5 [==============================] - 0s 17ms/step - loss: 15.5565 - mean_absolute_error: 2.6058 - val_loss: 31.6604 - val_mean_absolute_error: 3.0984\n",
            "Epoch 32/400\n",
            "5/5 [==============================] - 0s 22ms/step - loss: 15.2934 - mean_absolute_error: 2.4978 - val_loss: 30.9475 - val_mean_absolute_error: 3.0521\n",
            "Epoch 33/400\n",
            "5/5 [==============================] - 0s 22ms/step - loss: 15.0076 - mean_absolute_error: 2.5189 - val_loss: 30.3543 - val_mean_absolute_error: 3.0229\n",
            "Epoch 34/400\n",
            "5/5 [==============================] - 0s 19ms/step - loss: 14.7604 - mean_absolute_error: 2.4887 - val_loss: 29.6579 - val_mean_absolute_error: 2.9907\n",
            "Epoch 35/400\n",
            "5/5 [==============================] - 0s 21ms/step - loss: 14.5207 - mean_absolute_error: 2.4294 - val_loss: 28.9306 - val_mean_absolute_error: 2.9811\n",
            "Epoch 36/400\n",
            "5/5 [==============================] - 0s 20ms/step - loss: 14.3576 - mean_absolute_error: 2.4268 - val_loss: 28.2322 - val_mean_absolute_error: 2.9713\n",
            "Epoch 37/400\n",
            "5/5 [==============================] - 0s 20ms/step - loss: 14.1830 - mean_absolute_error: 2.5091 - val_loss: 27.7994 - val_mean_absolute_error: 2.9563\n",
            "Epoch 38/400\n",
            "5/5 [==============================] - 0s 20ms/step - loss: 13.9689 - mean_absolute_error: 2.4147 - val_loss: 27.7291 - val_mean_absolute_error: 2.9871\n",
            "Epoch 39/400\n",
            "5/5 [==============================] - 0s 21ms/step - loss: 13.7330 - mean_absolute_error: 2.3009 - val_loss: 27.7473 - val_mean_absolute_error: 3.0164\n",
            "Epoch 40/400\n",
            "5/5 [==============================] - 0s 24ms/step - loss: 13.5650 - mean_absolute_error: 2.5545 - val_loss: 27.5002 - val_mean_absolute_error: 3.0182\n",
            "Epoch 41/400\n",
            "5/5 [==============================] - 0s 17ms/step - loss: 13.4530 - mean_absolute_error: 2.4183 - val_loss: 27.1069 - val_mean_absolute_error: 2.9940\n",
            "Epoch 42/400\n",
            "5/5 [==============================] - 0s 14ms/step - loss: 13.2628 - mean_absolute_error: 2.4489 - val_loss: 27.0129 - val_mean_absolute_error: 2.9661\n",
            "Epoch 43/400\n",
            "5/5 [==============================] - 0s 15ms/step - loss: 13.0811 - mean_absolute_error: 2.3553 - val_loss: 26.7553 - val_mean_absolute_error: 2.9547\n",
            "Epoch 44/400\n",
            "5/5 [==============================] - 0s 11ms/step - loss: 12.9288 - mean_absolute_error: 2.3358 - val_loss: 26.7139 - val_mean_absolute_error: 2.9497\n",
            "Epoch 45/400\n",
            "5/5 [==============================] - 0s 12ms/step - loss: 12.8545 - mean_absolute_error: 2.3721 - val_loss: 26.4880 - val_mean_absolute_error: 2.9449\n",
            "Epoch 46/400\n",
            "5/5 [==============================] - 0s 11ms/step - loss: 12.7000 - mean_absolute_error: 2.3018 - val_loss: 26.0742 - val_mean_absolute_error: 2.9745\n",
            "Epoch 47/400\n",
            "5/5 [==============================] - 0s 16ms/step - loss: 12.5062 - mean_absolute_error: 2.3236 - val_loss: 25.8501 - val_mean_absolute_error: 3.0084\n",
            "Epoch 48/400\n",
            "5/5 [==============================] - 0s 15ms/step - loss: 12.4419 - mean_absolute_error: 2.3236 - val_loss: 25.7499 - val_mean_absolute_error: 3.0422\n",
            "Epoch 49/400\n",
            "5/5 [==============================] - 0s 11ms/step - loss: 12.4086 - mean_absolute_error: 2.3081 - val_loss: 25.1420 - val_mean_absolute_error: 3.0073\n",
            "Epoch 50/400\n",
            "5/5 [==============================] - 0s 15ms/step - loss: 12.2846 - mean_absolute_error: 2.2697 - val_loss: 24.8954 - val_mean_absolute_error: 2.9591\n",
            "Epoch 51/400\n",
            "5/5 [==============================] - 0s 12ms/step - loss: 12.1331 - mean_absolute_error: 2.3027 - val_loss: 24.9006 - val_mean_absolute_error: 2.9374\n",
            "Epoch 52/400\n",
            "5/5 [==============================] - 0s 15ms/step - loss: 12.0311 - mean_absolute_error: 2.3813 - val_loss: 25.2272 - val_mean_absolute_error: 2.9835\n",
            "Epoch 53/400\n",
            "5/5 [==============================] - 0s 15ms/step - loss: 12.0001 - mean_absolute_error: 2.5053 - val_loss: 25.3715 - val_mean_absolute_error: 2.9862\n",
            "Epoch 54/400\n",
            "5/5 [==============================] - 0s 15ms/step - loss: 11.7183 - mean_absolute_error: 2.2536 - val_loss: 24.5135 - val_mean_absolute_error: 2.9756\n",
            "Epoch 55/400\n",
            "5/5 [==============================] - 0s 16ms/step - loss: 11.7630 - mean_absolute_error: 2.5064 - val_loss: 24.0975 - val_mean_absolute_error: 2.9735\n",
            "Epoch 56/400\n",
            "5/5 [==============================] - 0s 15ms/step - loss: 11.5468 - mean_absolute_error: 2.4024 - val_loss: 24.3372 - val_mean_absolute_error: 2.9884\n",
            "Epoch 57/400\n",
            "5/5 [==============================] - 0s 16ms/step - loss: 11.4912 - mean_absolute_error: 2.3051 - val_loss: 24.9555 - val_mean_absolute_error: 3.0067\n",
            "Epoch 58/400\n",
            "5/5 [==============================] - 0s 15ms/step - loss: 11.4385 - mean_absolute_error: 2.3001 - val_loss: 25.1445 - val_mean_absolute_error: 3.0158\n",
            "Epoch 59/400\n",
            "5/5 [==============================] - 0s 12ms/step - loss: 11.3386 - mean_absolute_error: 2.2404 - val_loss: 24.7866 - val_mean_absolute_error: 2.9800\n",
            "Epoch 60/400\n",
            "5/5 [==============================] - 0s 15ms/step - loss: 11.1715 - mean_absolute_error: 2.2936 - val_loss: 24.2306 - val_mean_absolute_error: 2.9242\n",
            "Epoch 61/400\n",
            "5/5 [==============================] - 0s 15ms/step - loss: 11.1013 - mean_absolute_error: 2.4209 - val_loss: 23.8111 - val_mean_absolute_error: 2.9000\n",
            "Epoch 62/400\n",
            "5/5 [==============================] - 0s 10ms/step - loss: 11.0985 - mean_absolute_error: 2.1972 - val_loss: 23.5923 - val_mean_absolute_error: 2.9130\n",
            "Epoch 63/400\n",
            "5/5 [==============================] - 0s 10ms/step - loss: 10.9806 - mean_absolute_error: 2.1059 - val_loss: 23.7066 - val_mean_absolute_error: 2.9133\n",
            "Epoch 64/400\n",
            "5/5 [==============================] - 0s 16ms/step - loss: 10.8809 - mean_absolute_error: 2.2055 - val_loss: 23.9454 - val_mean_absolute_error: 2.9517\n",
            "Epoch 65/400\n",
            "5/5 [==============================] - 0s 11ms/step - loss: 10.7948 - mean_absolute_error: 2.2513 - val_loss: 24.0791 - val_mean_absolute_error: 2.9850\n",
            "Epoch 66/400\n",
            "5/5 [==============================] - 0s 13ms/step - loss: 10.6878 - mean_absolute_error: 2.1068 - val_loss: 23.8129 - val_mean_absolute_error: 2.9531\n",
            "Epoch 67/400\n",
            "5/5 [==============================] - 0s 17ms/step - loss: 10.6561 - mean_absolute_error: 2.2535 - val_loss: 23.3539 - val_mean_absolute_error: 2.8952\n",
            "Epoch 68/400\n",
            "5/5 [==============================] - 0s 11ms/step - loss: 10.6016 - mean_absolute_error: 2.0743 - val_loss: 23.2104 - val_mean_absolute_error: 2.8632\n",
            "Epoch 69/400\n",
            "5/5 [==============================] - 0s 12ms/step - loss: 10.4727 - mean_absolute_error: 2.1821 - val_loss: 23.4134 - val_mean_absolute_error: 2.8578\n",
            "Epoch 70/400\n",
            "5/5 [==============================] - 0s 11ms/step - loss: 10.4344 - mean_absolute_error: 2.0886 - val_loss: 23.8242 - val_mean_absolute_error: 2.8909\n",
            "Epoch 71/400\n",
            "5/5 [==============================] - 0s 15ms/step - loss: 10.4006 - mean_absolute_error: 2.0741 - val_loss: 23.7396 - val_mean_absolute_error: 2.8882\n",
            "Epoch 72/400\n",
            "5/5 [==============================] - 0s 16ms/step - loss: 10.3047 - mean_absolute_error: 2.2155 - val_loss: 23.7302 - val_mean_absolute_error: 2.9192\n",
            "Epoch 73/400\n",
            "5/5 [==============================] - 0s 15ms/step - loss: 10.2452 - mean_absolute_error: 2.0936 - val_loss: 23.3943 - val_mean_absolute_error: 2.8956\n",
            "Epoch 74/400\n",
            "5/5 [==============================] - 0s 12ms/step - loss: 10.1841 - mean_absolute_error: 2.1156 - val_loss: 22.9543 - val_mean_absolute_error: 2.8691\n",
            "Epoch 75/400\n",
            "5/5 [==============================] - 0s 12ms/step - loss: 10.1206 - mean_absolute_error: 2.2469 - val_loss: 23.0933 - val_mean_absolute_error: 2.9143\n",
            "Epoch 76/400\n",
            "5/5 [==============================] - 0s 11ms/step - loss: 10.0618 - mean_absolute_error: 2.0945 - val_loss: 23.1096 - val_mean_absolute_error: 2.8704\n",
            "Epoch 77/400\n",
            "5/5 [==============================] - 0s 12ms/step - loss: 10.0957 - mean_absolute_error: 2.1020 - val_loss: 23.1378 - val_mean_absolute_error: 2.8581\n",
            "Epoch 78/400\n",
            "5/5 [==============================] - 0s 16ms/step - loss: 10.0368 - mean_absolute_error: 2.0844 - val_loss: 22.8054 - val_mean_absolute_error: 2.8431\n",
            "Epoch 79/400\n",
            "5/5 [==============================] - 0s 12ms/step - loss: 9.9329 - mean_absolute_error: 2.0855 - val_loss: 22.9302 - val_mean_absolute_error: 2.8870\n",
            "Epoch 80/400\n",
            "5/5 [==============================] - 0s 11ms/step - loss: 9.7920 - mean_absolute_error: 2.0330 - val_loss: 22.7536 - val_mean_absolute_error: 2.8797\n",
            "Epoch 81/400\n",
            "5/5 [==============================] - 0s 16ms/step - loss: 9.7183 - mean_absolute_error: 2.0370 - val_loss: 22.5689 - val_mean_absolute_error: 2.9082\n",
            "Epoch 82/400\n",
            "5/5 [==============================] - 0s 16ms/step - loss: 9.6498 - mean_absolute_error: 2.0620 - val_loss: 22.4607 - val_mean_absolute_error: 2.9184\n",
            "Epoch 83/400\n",
            "5/5 [==============================] - 0s 15ms/step - loss: 9.6231 - mean_absolute_error: 2.0398 - val_loss: 22.2916 - val_mean_absolute_error: 2.9295\n",
            "Epoch 84/400\n",
            "5/5 [==============================] - 0s 15ms/step - loss: 9.5520 - mean_absolute_error: 1.9565 - val_loss: 22.6012 - val_mean_absolute_error: 2.9625\n",
            "Epoch 85/400\n",
            "5/5 [==============================] - 0s 12ms/step - loss: 9.5031 - mean_absolute_error: 2.1509 - val_loss: 22.7923 - val_mean_absolute_error: 2.9444\n",
            "Epoch 86/400\n",
            "5/5 [==============================] - 0s 14ms/step - loss: 9.4193 - mean_absolute_error: 2.0780 - val_loss: 22.8130 - val_mean_absolute_error: 2.9132\n",
            "Epoch 87/400\n",
            "5/5 [==============================] - 0s 13ms/step - loss: 9.3593 - mean_absolute_error: 1.9805 - val_loss: 22.7809 - val_mean_absolute_error: 2.8861\n",
            "Epoch 88/400\n",
            "5/5 [==============================] - 0s 14ms/step - loss: 9.3339 - mean_absolute_error: 2.0410 - val_loss: 22.8195 - val_mean_absolute_error: 2.8726\n",
            "Epoch 89/400\n",
            "5/5 [==============================] - 0s 12ms/step - loss: 9.2076 - mean_absolute_error: 2.0143 - val_loss: 22.6801 - val_mean_absolute_error: 2.8703\n",
            "Epoch 90/400\n",
            "5/5 [==============================] - 0s 17ms/step - loss: 9.1868 - mean_absolute_error: 1.9994 - val_loss: 22.5674 - val_mean_absolute_error: 2.8510\n",
            "Epoch 91/400\n",
            "5/5 [==============================] - 0s 17ms/step - loss: 9.1303 - mean_absolute_error: 2.0125 - val_loss: 22.7679 - val_mean_absolute_error: 2.8362\n",
            "Epoch 92/400\n",
            "5/5 [==============================] - 0s 17ms/step - loss: 9.0523 - mean_absolute_error: 2.0145 - val_loss: 22.4718 - val_mean_absolute_error: 2.8206\n",
            "Epoch 93/400\n",
            "5/5 [==============================] - 0s 17ms/step - loss: 9.0445 - mean_absolute_error: 1.9738 - val_loss: 22.2580 - val_mean_absolute_error: 2.8432\n",
            "Epoch 94/400\n",
            "5/5 [==============================] - 0s 12ms/step - loss: 8.9533 - mean_absolute_error: 1.9628 - val_loss: 21.9324 - val_mean_absolute_error: 2.7741\n",
            "Epoch 95/400\n",
            "5/5 [==============================] - 0s 13ms/step - loss: 8.9306 - mean_absolute_error: 2.0054 - val_loss: 21.6459 - val_mean_absolute_error: 2.6944\n",
            "Epoch 96/400\n",
            "5/5 [==============================] - 0s 17ms/step - loss: 8.9267 - mean_absolute_error: 1.9295 - val_loss: 21.7976 - val_mean_absolute_error: 2.6919\n",
            "Epoch 97/400\n",
            "5/5 [==============================] - 0s 16ms/step - loss: 8.7779 - mean_absolute_error: 2.0654 - val_loss: 21.9602 - val_mean_absolute_error: 2.7085\n",
            "Epoch 98/400\n",
            "5/5 [==============================] - 0s 18ms/step - loss: 8.7236 - mean_absolute_error: 2.0120 - val_loss: 21.7916 - val_mean_absolute_error: 2.7208\n",
            "Epoch 99/400\n",
            "5/5 [==============================] - 0s 13ms/step - loss: 8.6254 - mean_absolute_error: 1.9515 - val_loss: 21.9293 - val_mean_absolute_error: 2.7461\n",
            "Epoch 100/400\n",
            "5/5 [==============================] - 0s 20ms/step - loss: 8.5700 - mean_absolute_error: 1.9346 - val_loss: 22.1115 - val_mean_absolute_error: 2.7451\n",
            "Epoch 101/400\n",
            "5/5 [==============================] - 0s 12ms/step - loss: 8.5150 - mean_absolute_error: 1.9303 - val_loss: 22.4780 - val_mean_absolute_error: 2.8058\n",
            "Epoch 102/400\n",
            "5/5 [==============================] - 0s 13ms/step - loss: 8.5122 - mean_absolute_error: 1.8647 - val_loss: 22.5777 - val_mean_absolute_error: 2.8738\n",
            "Epoch 103/400\n",
            "5/5 [==============================] - 0s 16ms/step - loss: 8.4578 - mean_absolute_error: 1.8454 - val_loss: 22.3469 - val_mean_absolute_error: 2.8381\n",
            "Epoch 104/400\n",
            "5/5 [==============================] - 0s 17ms/step - loss: 8.3404 - mean_absolute_error: 1.8221 - val_loss: 22.1422 - val_mean_absolute_error: 2.7870\n",
            "Epoch 105/400\n",
            "5/5 [==============================] - 0s 13ms/step - loss: 8.2736 - mean_absolute_error: 2.0499 - val_loss: 22.0391 - val_mean_absolute_error: 2.7655\n",
            "Epoch 106/400\n",
            "5/5 [==============================] - 0s 12ms/step - loss: 8.2204 - mean_absolute_error: 1.8503 - val_loss: 22.3172 - val_mean_absolute_error: 2.7608\n",
            "Epoch 107/400\n",
            "5/5 [==============================] - 0s 16ms/step - loss: 8.1749 - mean_absolute_error: 1.7952 - val_loss: 22.2919 - val_mean_absolute_error: 2.7095\n",
            "Epoch 108/400\n",
            "5/5 [==============================] - 0s 11ms/step - loss: 8.1332 - mean_absolute_error: 1.8665 - val_loss: 22.1185 - val_mean_absolute_error: 2.6914\n",
            "Epoch 109/400\n",
            "5/5 [==============================] - 0s 12ms/step - loss: 8.0623 - mean_absolute_error: 1.9492 - val_loss: 22.1185 - val_mean_absolute_error: 2.6967\n",
            "Epoch 110/400\n",
            "5/5 [==============================] - 0s 11ms/step - loss: 8.0697 - mean_absolute_error: 1.9808 - val_loss: 22.2969 - val_mean_absolute_error: 2.7156\n",
            "Epoch 111/400\n",
            "5/5 [==============================] - 0s 15ms/step - loss: 7.9520 - mean_absolute_error: 1.9409 - val_loss: 22.7766 - val_mean_absolute_error: 2.7727\n",
            "Epoch 112/400\n",
            "5/5 [==============================] - 0s 15ms/step - loss: 8.0064 - mean_absolute_error: 1.9134 - val_loss: 22.8837 - val_mean_absolute_error: 2.7315\n",
            "Epoch 113/400\n",
            "5/5 [==============================] - 0s 15ms/step - loss: 7.8539 - mean_absolute_error: 1.8376 - val_loss: 22.6186 - val_mean_absolute_error: 2.6861\n",
            "Epoch 114/400\n",
            "5/5 [==============================] - 0s 15ms/step - loss: 7.8210 - mean_absolute_error: 1.7975 - val_loss: 21.9294 - val_mean_absolute_error: 2.6454\n",
            "Epoch 115/400\n",
            "5/5 [==============================] - 0s 16ms/step - loss: 7.7927 - mean_absolute_error: 1.8717 - val_loss: 21.7360 - val_mean_absolute_error: 2.6844\n",
            "Epoch 116/400\n",
            "5/5 [==============================] - 0s 11ms/step - loss: 7.7031 - mean_absolute_error: 1.7977 - val_loss: 21.8875 - val_mean_absolute_error: 2.7112\n",
            "Epoch 117/400\n",
            "5/5 [==============================] - 0s 15ms/step - loss: 7.6517 - mean_absolute_error: 1.7976 - val_loss: 22.0523 - val_mean_absolute_error: 2.6916\n",
            "Epoch 118/400\n",
            "5/5 [==============================] - 0s 15ms/step - loss: 7.5902 - mean_absolute_error: 1.8141 - val_loss: 22.3503 - val_mean_absolute_error: 2.6968\n",
            "Epoch 119/400\n",
            "5/5 [==============================] - 0s 15ms/step - loss: 7.5547 - mean_absolute_error: 1.8596 - val_loss: 22.2809 - val_mean_absolute_error: 2.7172\n",
            "Epoch 120/400\n",
            "5/5 [==============================] - 0s 15ms/step - loss: 7.4992 - mean_absolute_error: 1.7434 - val_loss: 22.1701 - val_mean_absolute_error: 2.6951\n",
            "Epoch 121/400\n",
            "5/5 [==============================] - 0s 11ms/step - loss: 7.4521 - mean_absolute_error: 1.8745 - val_loss: 22.0821 - val_mean_absolute_error: 2.7005\n",
            "Epoch 122/400\n",
            "5/5 [==============================] - 0s 16ms/step - loss: 7.3920 - mean_absolute_error: 1.8328 - val_loss: 22.2900 - val_mean_absolute_error: 2.7155\n",
            "Epoch 123/400\n",
            "5/5 [==============================] - 0s 15ms/step - loss: 7.4310 - mean_absolute_error: 1.8984 - val_loss: 22.1910 - val_mean_absolute_error: 2.6659\n",
            "Epoch 124/400\n",
            "5/5 [==============================] - 0s 12ms/step - loss: 7.3172 - mean_absolute_error: 1.8556 - val_loss: 21.7197 - val_mean_absolute_error: 2.6134\n",
            "Epoch 125/400\n",
            "5/5 [==============================] - 0s 15ms/step - loss: 7.3301 - mean_absolute_error: 1.7403 - val_loss: 21.8507 - val_mean_absolute_error: 2.6375\n",
            "Epoch 125: early stopping\n"
          ]
        }
      ],
      "source": [
        "history_reg = model_reg.fit(X_train, y_train, validation_split=0.2, batch_size=BATCH_SIZE, epochs=NUM_EPOCHS, verbose=1, callbacks=early_stop)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "xWFUiaz-uDND",
        "outputId": "f92e27a9-b720-41b5-e907-9bf2a8a34564"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/400\n",
            "5/5 [==============================] - 1s 57ms/step - loss: 566.7428 - mean_absolute_error: 21.7932 - val_loss: 574.3729 - val_mean_absolute_error: 20.6601\n",
            "Epoch 2/400\n",
            "5/5 [==============================] - 0s 16ms/step - loss: 522.0514 - mean_absolute_error: 21.3671 - val_loss: 527.1069 - val_mean_absolute_error: 19.6217\n",
            "Epoch 3/400\n",
            "5/5 [==============================] - 0s 13ms/step - loss: 475.0437 - mean_absolute_error: 19.9601 - val_loss: 475.8532 - val_mean_absolute_error: 18.4499\n",
            "Epoch 4/400\n",
            "5/5 [==============================] - 0s 15ms/step - loss: 421.1752 - mean_absolute_error: 18.4244 - val_loss: 417.1227 - val_mean_absolute_error: 17.0438\n",
            "Epoch 5/400\n",
            "5/5 [==============================] - 0s 11ms/step - loss: 362.6312 - mean_absolute_error: 17.4229 - val_loss: 351.4537 - val_mean_absolute_error: 15.3652\n",
            "Epoch 6/400\n",
            "5/5 [==============================] - 0s 15ms/step - loss: 293.1451 - mean_absolute_error: 15.1112 - val_loss: 282.2778 - val_mean_absolute_error: 13.3988\n",
            "Epoch 7/400\n",
            "5/5 [==============================] - 0s 15ms/step - loss: 238.4551 - mean_absolute_error: 12.8748 - val_loss: 214.5041 - val_mean_absolute_error: 11.1813\n",
            "Epoch 8/400\n",
            "5/5 [==============================] - 0s 18ms/step - loss: 166.9792 - mean_absolute_error: 10.8683 - val_loss: 156.5199 - val_mean_absolute_error: 8.7928\n",
            "Epoch 9/400\n",
            "5/5 [==============================] - 0s 15ms/step - loss: 120.3847 - mean_absolute_error: 8.0569 - val_loss: 115.0189 - val_mean_absolute_error: 6.5844\n",
            "Epoch 10/400\n",
            "5/5 [==============================] - 0s 12ms/step - loss: 89.8045 - mean_absolute_error: 7.1585 - val_loss: 91.9708 - val_mean_absolute_error: 5.3688\n",
            "Epoch 11/400\n",
            "5/5 [==============================] - 0s 16ms/step - loss: 70.9368 - mean_absolute_error: 6.3861 - val_loss: 79.7850 - val_mean_absolute_error: 5.2622\n",
            "Epoch 12/400\n",
            "5/5 [==============================] - 0s 15ms/step - loss: 62.6245 - mean_absolute_error: 5.9442 - val_loss: 70.2931 - val_mean_absolute_error: 4.9185\n",
            "Epoch 13/400\n",
            "5/5 [==============================] - 0s 12ms/step - loss: 46.7136 - mean_absolute_error: 5.0791 - val_loss: 60.9954 - val_mean_absolute_error: 4.2885\n",
            "Epoch 14/400\n",
            "5/5 [==============================] - 0s 14ms/step - loss: 43.4384 - mean_absolute_error: 5.1898 - val_loss: 53.3254 - val_mean_absolute_error: 3.6013\n",
            "Epoch 15/400\n",
            "5/5 [==============================] - 0s 11ms/step - loss: 35.8953 - mean_absolute_error: 4.3880 - val_loss: 47.8540 - val_mean_absolute_error: 3.4495\n",
            "Epoch 16/400\n",
            "5/5 [==============================] - 0s 11ms/step - loss: 31.4587 - mean_absolute_error: 4.2464 - val_loss: 44.6053 - val_mean_absolute_error: 3.5511\n",
            "Epoch 17/400\n",
            "5/5 [==============================] - 0s 15ms/step - loss: 29.7088 - mean_absolute_error: 4.2687 - val_loss: 42.2527 - val_mean_absolute_error: 3.6198\n",
            "Epoch 18/400\n",
            "5/5 [==============================] - 0s 12ms/step - loss: 26.1846 - mean_absolute_error: 3.8344 - val_loss: 40.0183 - val_mean_absolute_error: 3.6164\n",
            "Epoch 19/400\n",
            "5/5 [==============================] - 0s 11ms/step - loss: 27.6931 - mean_absolute_error: 3.7708 - val_loss: 38.6552 - val_mean_absolute_error: 3.6054\n",
            "Epoch 20/400\n",
            "5/5 [==============================] - 0s 15ms/step - loss: 25.0450 - mean_absolute_error: 3.6341 - val_loss: 37.4354 - val_mean_absolute_error: 3.5344\n",
            "Epoch 21/400\n",
            "5/5 [==============================] - 0s 12ms/step - loss: 25.0599 - mean_absolute_error: 3.8605 - val_loss: 36.2799 - val_mean_absolute_error: 3.4247\n",
            "Epoch 22/400\n",
            "5/5 [==============================] - 0s 11ms/step - loss: 21.5369 - mean_absolute_error: 3.4795 - val_loss: 35.3531 - val_mean_absolute_error: 3.3454\n",
            "Epoch 23/400\n",
            "5/5 [==============================] - 0s 11ms/step - loss: 24.7011 - mean_absolute_error: 3.5864 - val_loss: 34.5208 - val_mean_absolute_error: 3.3825\n",
            "Epoch 24/400\n",
            "5/5 [==============================] - 0s 14ms/step - loss: 23.0649 - mean_absolute_error: 3.7440 - val_loss: 33.8253 - val_mean_absolute_error: 3.4392\n",
            "Epoch 25/400\n",
            "5/5 [==============================] - 0s 12ms/step - loss: 26.2715 - mean_absolute_error: 3.6907 - val_loss: 32.7984 - val_mean_absolute_error: 3.4227\n",
            "Epoch 26/400\n",
            "5/5 [==============================] - 0s 15ms/step - loss: 22.1662 - mean_absolute_error: 3.5596 - val_loss: 31.9655 - val_mean_absolute_error: 3.3834\n",
            "Epoch 27/400\n",
            "5/5 [==============================] - 0s 15ms/step - loss: 20.8628 - mean_absolute_error: 3.7250 - val_loss: 31.3370 - val_mean_absolute_error: 3.3150\n",
            "Epoch 28/400\n",
            "5/5 [==============================] - 0s 11ms/step - loss: 20.8898 - mean_absolute_error: 3.5719 - val_loss: 31.1595 - val_mean_absolute_error: 3.3067\n",
            "Epoch 29/400\n",
            "5/5 [==============================] - 0s 11ms/step - loss: 22.4436 - mean_absolute_error: 3.8223 - val_loss: 30.4527 - val_mean_absolute_error: 3.2175\n",
            "Epoch 30/400\n",
            "5/5 [==============================] - 0s 15ms/step - loss: 20.6096 - mean_absolute_error: 3.3560 - val_loss: 29.9204 - val_mean_absolute_error: 3.1138\n",
            "Epoch 31/400\n",
            "5/5 [==============================] - 0s 13ms/step - loss: 19.1350 - mean_absolute_error: 3.3710 - val_loss: 29.4054 - val_mean_absolute_error: 3.0877\n",
            "Epoch 32/400\n",
            "5/5 [==============================] - 0s 15ms/step - loss: 21.4306 - mean_absolute_error: 3.4330 - val_loss: 28.6941 - val_mean_absolute_error: 3.0996\n",
            "Epoch 33/400\n",
            "5/5 [==============================] - 0s 15ms/step - loss: 20.3376 - mean_absolute_error: 3.1561 - val_loss: 28.1042 - val_mean_absolute_error: 3.1912\n",
            "Epoch 34/400\n",
            "5/5 [==============================] - 0s 15ms/step - loss: 21.7285 - mean_absolute_error: 3.4588 - val_loss: 27.7421 - val_mean_absolute_error: 3.2248\n",
            "Epoch 35/400\n",
            "5/5 [==============================] - 0s 13ms/step - loss: 22.0444 - mean_absolute_error: 3.4323 - val_loss: 27.4864 - val_mean_absolute_error: 3.2304\n",
            "Epoch 36/400\n",
            "5/5 [==============================] - 0s 11ms/step - loss: 18.9615 - mean_absolute_error: 3.2039 - val_loss: 26.8633 - val_mean_absolute_error: 3.1172\n",
            "Epoch 37/400\n",
            "5/5 [==============================] - 0s 21ms/step - loss: 21.5968 - mean_absolute_error: 3.6335 - val_loss: 26.3195 - val_mean_absolute_error: 3.0430\n",
            "Epoch 38/400\n",
            "5/5 [==============================] - 0s 20ms/step - loss: 21.2679 - mean_absolute_error: 3.4169 - val_loss: 25.8546 - val_mean_absolute_error: 3.0458\n",
            "Epoch 39/400\n",
            "5/5 [==============================] - 0s 21ms/step - loss: 18.4590 - mean_absolute_error: 3.2386 - val_loss: 25.5205 - val_mean_absolute_error: 3.0244\n",
            "Epoch 40/400\n",
            "5/5 [==============================] - 0s 16ms/step - loss: 17.9184 - mean_absolute_error: 3.2009 - val_loss: 25.2810 - val_mean_absolute_error: 3.0197\n",
            "Epoch 41/400\n",
            "5/5 [==============================] - 0s 19ms/step - loss: 19.7411 - mean_absolute_error: 3.2326 - val_loss: 25.1056 - val_mean_absolute_error: 3.0666\n",
            "Epoch 42/400\n",
            "5/5 [==============================] - 0s 16ms/step - loss: 15.9398 - mean_absolute_error: 2.9608 - val_loss: 25.0228 - val_mean_absolute_error: 3.0633\n",
            "Epoch 43/400\n",
            "5/5 [==============================] - 0s 20ms/step - loss: 20.2203 - mean_absolute_error: 3.0024 - val_loss: 24.7216 - val_mean_absolute_error: 3.0458\n",
            "Epoch 44/400\n",
            "5/5 [==============================] - 0s 19ms/step - loss: 20.5428 - mean_absolute_error: 3.2601 - val_loss: 23.9611 - val_mean_absolute_error: 2.9389\n",
            "Epoch 45/400\n",
            "5/5 [==============================] - 0s 18ms/step - loss: 17.0091 - mean_absolute_error: 3.0606 - val_loss: 23.4676 - val_mean_absolute_error: 2.9127\n",
            "Epoch 46/400\n",
            "5/5 [==============================] - 0s 17ms/step - loss: 18.8244 - mean_absolute_error: 3.1626 - val_loss: 23.1880 - val_mean_absolute_error: 2.9433\n",
            "Epoch 47/400\n",
            "5/5 [==============================] - 0s 20ms/step - loss: 19.7038 - mean_absolute_error: 3.0700 - val_loss: 22.8436 - val_mean_absolute_error: 2.9244\n",
            "Epoch 48/400\n",
            "5/5 [==============================] - 0s 15ms/step - loss: 18.1029 - mean_absolute_error: 3.2213 - val_loss: 22.8219 - val_mean_absolute_error: 2.9519\n",
            "Epoch 49/400\n",
            "5/5 [==============================] - 0s 15ms/step - loss: 18.1569 - mean_absolute_error: 3.0853 - val_loss: 22.8769 - val_mean_absolute_error: 2.9621\n",
            "Epoch 50/400\n",
            "5/5 [==============================] - 0s 19ms/step - loss: 17.3872 - mean_absolute_error: 3.0682 - val_loss: 22.6557 - val_mean_absolute_error: 2.9223\n",
            "Epoch 51/400\n",
            "5/5 [==============================] - 0s 21ms/step - loss: 15.4280 - mean_absolute_error: 2.9328 - val_loss: 22.0935 - val_mean_absolute_error: 2.8588\n",
            "Epoch 52/400\n",
            "5/5 [==============================] - 0s 15ms/step - loss: 16.3084 - mean_absolute_error: 3.1920 - val_loss: 21.8228 - val_mean_absolute_error: 2.8195\n",
            "Epoch 53/400\n",
            "5/5 [==============================] - 0s 18ms/step - loss: 16.3975 - mean_absolute_error: 2.9660 - val_loss: 21.6602 - val_mean_absolute_error: 2.8623\n",
            "Epoch 54/400\n",
            "5/5 [==============================] - 0s 18ms/step - loss: 18.8100 - mean_absolute_error: 3.0493 - val_loss: 22.0166 - val_mean_absolute_error: 2.9742\n",
            "Epoch 55/400\n",
            "5/5 [==============================] - 0s 19ms/step - loss: 18.4474 - mean_absolute_error: 3.4218 - val_loss: 22.2536 - val_mean_absolute_error: 3.0133\n",
            "Epoch 56/400\n",
            "5/5 [==============================] - 0s 20ms/step - loss: 16.7059 - mean_absolute_error: 2.8594 - val_loss: 21.6882 - val_mean_absolute_error: 2.9726\n",
            "Epoch 57/400\n",
            "5/5 [==============================] - 0s 19ms/step - loss: 15.3406 - mean_absolute_error: 2.8378 - val_loss: 21.2156 - val_mean_absolute_error: 2.9246\n",
            "Epoch 58/400\n",
            "5/5 [==============================] - 0s 21ms/step - loss: 16.6112 - mean_absolute_error: 2.8861 - val_loss: 21.0367 - val_mean_absolute_error: 2.8985\n",
            "Epoch 59/400\n",
            "5/5 [==============================] - 0s 19ms/step - loss: 15.8656 - mean_absolute_error: 2.9477 - val_loss: 20.8723 - val_mean_absolute_error: 2.8697\n",
            "Epoch 60/400\n",
            "5/5 [==============================] - 0s 19ms/step - loss: 17.5503 - mean_absolute_error: 3.2068 - val_loss: 20.7469 - val_mean_absolute_error: 2.8459\n",
            "Epoch 61/400\n",
            "5/5 [==============================] - 0s 21ms/step - loss: 15.9717 - mean_absolute_error: 2.9591 - val_loss: 21.3892 - val_mean_absolute_error: 2.9301\n",
            "Epoch 62/400\n",
            "5/5 [==============================] - 0s 19ms/step - loss: 15.5932 - mean_absolute_error: 2.9507 - val_loss: 22.1521 - val_mean_absolute_error: 2.9803\n",
            "Epoch 63/400\n",
            "5/5 [==============================] - 0s 23ms/step - loss: 14.4505 - mean_absolute_error: 2.5566 - val_loss: 22.2831 - val_mean_absolute_error: 2.9154\n",
            "Epoch 64/400\n",
            "5/5 [==============================] - 0s 20ms/step - loss: 15.6869 - mean_absolute_error: 2.9486 - val_loss: 21.8692 - val_mean_absolute_error: 2.8938\n",
            "Epoch 65/400\n",
            "5/5 [==============================] - 0s 11ms/step - loss: 15.8172 - mean_absolute_error: 2.8639 - val_loss: 21.0043 - val_mean_absolute_error: 2.7891\n",
            "Epoch 66/400\n",
            "5/5 [==============================] - 0s 11ms/step - loss: 14.4734 - mean_absolute_error: 2.8649 - val_loss: 20.7080 - val_mean_absolute_error: 2.7663\n",
            "Epoch 67/400\n",
            "5/5 [==============================] - 0s 11ms/step - loss: 14.7559 - mean_absolute_error: 2.6723 - val_loss: 20.7313 - val_mean_absolute_error: 2.8210\n",
            "Epoch 68/400\n",
            "5/5 [==============================] - 0s 12ms/step - loss: 14.8863 - mean_absolute_error: 2.8499 - val_loss: 20.7115 - val_mean_absolute_error: 2.8354\n",
            "Epoch 69/400\n",
            "5/5 [==============================] - 0s 10ms/step - loss: 14.2191 - mean_absolute_error: 2.7435 - val_loss: 20.7857 - val_mean_absolute_error: 2.8882\n",
            "Epoch 70/400\n",
            "5/5 [==============================] - 0s 11ms/step - loss: 14.6336 - mean_absolute_error: 2.7532 - val_loss: 20.6109 - val_mean_absolute_error: 2.9082\n",
            "Epoch 71/400\n",
            "5/5 [==============================] - 0s 11ms/step - loss: 16.9261 - mean_absolute_error: 3.3050 - val_loss: 20.2698 - val_mean_absolute_error: 2.8928\n",
            "Epoch 72/400\n",
            "5/5 [==============================] - 0s 11ms/step - loss: 16.8912 - mean_absolute_error: 2.8682 - val_loss: 20.0055 - val_mean_absolute_error: 2.8480\n",
            "Epoch 73/400\n",
            "5/5 [==============================] - 0s 15ms/step - loss: 13.8975 - mean_absolute_error: 3.0224 - val_loss: 19.9023 - val_mean_absolute_error: 2.8171\n",
            "Epoch 74/400\n",
            "5/5 [==============================] - 0s 15ms/step - loss: 15.9419 - mean_absolute_error: 2.8422 - val_loss: 20.0133 - val_mean_absolute_error: 2.8528\n",
            "Epoch 75/400\n",
            "5/5 [==============================] - 0s 16ms/step - loss: 14.1153 - mean_absolute_error: 2.6681 - val_loss: 20.5373 - val_mean_absolute_error: 2.9386\n",
            "Epoch 76/400\n",
            "5/5 [==============================] - 0s 12ms/step - loss: 17.0244 - mean_absolute_error: 3.2023 - val_loss: 20.2459 - val_mean_absolute_error: 2.9008\n",
            "Epoch 77/400\n",
            "5/5 [==============================] - 0s 16ms/step - loss: 14.4614 - mean_absolute_error: 2.8576 - val_loss: 19.8045 - val_mean_absolute_error: 2.8093\n",
            "Epoch 78/400\n",
            "5/5 [==============================] - 0s 12ms/step - loss: 14.0634 - mean_absolute_error: 2.7826 - val_loss: 20.2625 - val_mean_absolute_error: 2.8669\n",
            "Epoch 79/400\n",
            "5/5 [==============================] - 0s 18ms/step - loss: 15.8822 - mean_absolute_error: 3.2183 - val_loss: 20.5960 - val_mean_absolute_error: 2.9409\n",
            "Epoch 80/400\n",
            "5/5 [==============================] - 0s 11ms/step - loss: 15.9634 - mean_absolute_error: 2.8112 - val_loss: 19.7054 - val_mean_absolute_error: 2.8607\n",
            "Epoch 81/400\n",
            "5/5 [==============================] - 0s 11ms/step - loss: 15.1555 - mean_absolute_error: 2.7934 - val_loss: 19.0671 - val_mean_absolute_error: 2.8226\n",
            "Epoch 82/400\n",
            "5/5 [==============================] - 0s 15ms/step - loss: 15.5798 - mean_absolute_error: 2.7103 - val_loss: 19.0533 - val_mean_absolute_error: 2.8444\n",
            "Epoch 83/400\n",
            "5/5 [==============================] - 0s 12ms/step - loss: 16.1853 - mean_absolute_error: 3.0730 - val_loss: 19.4923 - val_mean_absolute_error: 2.9151\n",
            "Epoch 84/400\n",
            "5/5 [==============================] - 0s 11ms/step - loss: 17.3038 - mean_absolute_error: 3.3151 - val_loss: 19.8284 - val_mean_absolute_error: 2.9135\n",
            "Epoch 85/400\n",
            "5/5 [==============================] - 0s 15ms/step - loss: 12.9801 - mean_absolute_error: 2.8233 - val_loss: 19.1869 - val_mean_absolute_error: 2.8280\n",
            "Epoch 86/400\n",
            "5/5 [==============================] - 0s 12ms/step - loss: 13.2085 - mean_absolute_error: 2.8244 - val_loss: 18.9195 - val_mean_absolute_error: 2.8243\n",
            "Epoch 87/400\n",
            "5/5 [==============================] - 0s 12ms/step - loss: 15.4102 - mean_absolute_error: 3.0836 - val_loss: 18.9781 - val_mean_absolute_error: 2.8917\n",
            "Epoch 88/400\n",
            "5/5 [==============================] - 0s 15ms/step - loss: 16.9121 - mean_absolute_error: 2.7598 - val_loss: 19.6907 - val_mean_absolute_error: 2.9472\n",
            "Epoch 89/400\n",
            "5/5 [==============================] - 0s 16ms/step - loss: 16.3020 - mean_absolute_error: 3.0959 - val_loss: 20.3858 - val_mean_absolute_error: 2.9693\n",
            "Epoch 90/400\n",
            "5/5 [==============================] - 0s 11ms/step - loss: 14.2250 - mean_absolute_error: 2.6686 - val_loss: 20.1387 - val_mean_absolute_error: 2.8706\n",
            "Epoch 91/400\n",
            "5/5 [==============================] - 0s 16ms/step - loss: 14.7337 - mean_absolute_error: 2.9807 - val_loss: 19.8635 - val_mean_absolute_error: 2.8042\n",
            "Epoch 92/400\n",
            "5/5 [==============================] - 0s 11ms/step - loss: 14.8041 - mean_absolute_error: 2.6386 - val_loss: 20.4578 - val_mean_absolute_error: 2.8696\n",
            "Epoch 93/400\n",
            "5/5 [==============================] - 0s 12ms/step - loss: 12.7415 - mean_absolute_error: 2.7756 - val_loss: 20.3262 - val_mean_absolute_error: 2.8511\n",
            "Epoch 94/400\n",
            "5/5 [==============================] - 0s 13ms/step - loss: 14.8096 - mean_absolute_error: 2.9852 - val_loss: 20.2230 - val_mean_absolute_error: 2.8323\n",
            "Epoch 95/400\n",
            "5/5 [==============================] - 0s 16ms/step - loss: 13.8162 - mean_absolute_error: 2.9195 - val_loss: 19.8678 - val_mean_absolute_error: 2.7565\n",
            "Epoch 96/400\n",
            "5/5 [==============================] - 0s 16ms/step - loss: 15.7202 - mean_absolute_error: 2.9367 - val_loss: 20.4453 - val_mean_absolute_error: 2.7654\n",
            "Epoch 97/400\n",
            "5/5 [==============================] - 0s 11ms/step - loss: 13.0580 - mean_absolute_error: 2.6189 - val_loss: 20.3953 - val_mean_absolute_error: 2.7955\n",
            "Epoch 98/400\n",
            "5/5 [==============================] - 0s 12ms/step - loss: 12.4226 - mean_absolute_error: 2.6310 - val_loss: 20.3381 - val_mean_absolute_error: 2.8584\n",
            "Epoch 99/400\n",
            "5/5 [==============================] - 0s 11ms/step - loss: 14.3626 - mean_absolute_error: 2.8561 - val_loss: 20.7541 - val_mean_absolute_error: 2.9532\n",
            "Epoch 100/400\n",
            "5/5 [==============================] - 0s 12ms/step - loss: 15.1437 - mean_absolute_error: 3.0234 - val_loss: 20.1151 - val_mean_absolute_error: 2.9127\n",
            "Epoch 101/400\n",
            "5/5 [==============================] - 0s 17ms/step - loss: 14.1648 - mean_absolute_error: 2.8289 - val_loss: 19.4255 - val_mean_absolute_error: 2.8635\n",
            "Epoch 102/400\n",
            "5/5 [==============================] - 0s 12ms/step - loss: 13.8472 - mean_absolute_error: 2.7772 - val_loss: 19.3025 - val_mean_absolute_error: 2.8573\n",
            "Epoch 103/400\n",
            "5/5 [==============================] - 0s 12ms/step - loss: 12.2718 - mean_absolute_error: 2.8207 - val_loss: 19.4839 - val_mean_absolute_error: 2.8881\n",
            "Epoch 104/400\n",
            "5/5 [==============================] - 0s 17ms/step - loss: 13.2255 - mean_absolute_error: 2.7851 - val_loss: 19.0849 - val_mean_absolute_error: 2.8578\n",
            "Epoch 105/400\n",
            "5/5 [==============================] - 0s 15ms/step - loss: 14.3434 - mean_absolute_error: 2.9170 - val_loss: 19.0447 - val_mean_absolute_error: 2.8989\n",
            "Epoch 106/400\n",
            "5/5 [==============================] - 0s 13ms/step - loss: 15.3892 - mean_absolute_error: 2.6548 - val_loss: 19.0389 - val_mean_absolute_error: 2.8777\n",
            "Epoch 107/400\n",
            "5/5 [==============================] - 0s 16ms/step - loss: 16.1395 - mean_absolute_error: 2.9511 - val_loss: 19.0256 - val_mean_absolute_error: 2.8701\n",
            "Epoch 108/400\n",
            "5/5 [==============================] - 0s 12ms/step - loss: 13.7246 - mean_absolute_error: 2.7199 - val_loss: 18.8100 - val_mean_absolute_error: 2.7909\n",
            "Epoch 109/400\n",
            "5/5 [==============================] - 0s 16ms/step - loss: 12.8699 - mean_absolute_error: 2.6594 - val_loss: 18.6634 - val_mean_absolute_error: 2.7462\n",
            "Epoch 110/400\n",
            "5/5 [==============================] - 0s 14ms/step - loss: 12.4680 - mean_absolute_error: 2.6929 - val_loss: 19.3860 - val_mean_absolute_error: 2.8254\n",
            "Epoch 111/400\n",
            "5/5 [==============================] - 0s 15ms/step - loss: 12.8694 - mean_absolute_error: 2.6232 - val_loss: 20.5454 - val_mean_absolute_error: 2.9005\n",
            "Epoch 112/400\n",
            "5/5 [==============================] - 0s 12ms/step - loss: 12.2395 - mean_absolute_error: 2.6766 - val_loss: 20.0624 - val_mean_absolute_error: 2.8100\n",
            "Epoch 113/400\n",
            "5/5 [==============================] - 0s 11ms/step - loss: 13.5380 - mean_absolute_error: 2.8836 - val_loss: 19.3071 - val_mean_absolute_error: 2.7129\n",
            "Epoch 114/400\n",
            "5/5 [==============================] - 0s 16ms/step - loss: 14.3768 - mean_absolute_error: 2.8284 - val_loss: 19.2991 - val_mean_absolute_error: 2.6657\n",
            "Epoch 115/400\n",
            "5/5 [==============================] - 0s 12ms/step - loss: 12.8692 - mean_absolute_error: 2.8052 - val_loss: 18.9688 - val_mean_absolute_error: 2.6157\n",
            "Epoch 116/400\n",
            "5/5 [==============================] - 0s 16ms/step - loss: 14.1401 - mean_absolute_error: 2.6703 - val_loss: 18.6337 - val_mean_absolute_error: 2.6366\n",
            "Epoch 117/400\n",
            "5/5 [==============================] - 0s 16ms/step - loss: 14.5492 - mean_absolute_error: 2.8116 - val_loss: 18.6161 - val_mean_absolute_error: 2.7222\n",
            "Epoch 118/400\n",
            "5/5 [==============================] - 0s 14ms/step - loss: 13.7335 - mean_absolute_error: 2.8435 - val_loss: 19.0933 - val_mean_absolute_error: 2.8396\n",
            "Epoch 119/400\n",
            "5/5 [==============================] - 0s 15ms/step - loss: 15.6893 - mean_absolute_error: 2.9875 - val_loss: 19.1688 - val_mean_absolute_error: 2.8506\n",
            "Epoch 120/400\n",
            "5/5 [==============================] - 0s 16ms/step - loss: 12.2895 - mean_absolute_error: 2.6824 - val_loss: 18.5726 - val_mean_absolute_error: 2.7510\n",
            "Epoch 121/400\n",
            "5/5 [==============================] - 0s 16ms/step - loss: 13.4971 - mean_absolute_error: 2.7436 - val_loss: 18.5318 - val_mean_absolute_error: 2.7383\n",
            "Epoch 122/400\n",
            "5/5 [==============================] - 0s 16ms/step - loss: 12.8957 - mean_absolute_error: 2.6860 - val_loss: 18.5723 - val_mean_absolute_error: 2.8026\n",
            "Epoch 123/400\n",
            "5/5 [==============================] - 0s 16ms/step - loss: 11.5910 - mean_absolute_error: 2.6455 - val_loss: 18.8349 - val_mean_absolute_error: 2.8851\n",
            "Epoch 124/400\n",
            "5/5 [==============================] - 0s 17ms/step - loss: 13.8239 - mean_absolute_error: 2.7994 - val_loss: 18.8795 - val_mean_absolute_error: 2.8616\n",
            "Epoch 125/400\n",
            "5/5 [==============================] - 0s 12ms/step - loss: 14.5766 - mean_absolute_error: 2.9813 - val_loss: 18.8172 - val_mean_absolute_error: 2.8411\n",
            "Epoch 126/400\n",
            "5/5 [==============================] - 0s 12ms/step - loss: 12.8128 - mean_absolute_error: 2.7844 - val_loss: 18.3204 - val_mean_absolute_error: 2.7162\n",
            "Epoch 127/400\n",
            "5/5 [==============================] - 0s 11ms/step - loss: 14.5794 - mean_absolute_error: 2.8319 - val_loss: 18.2512 - val_mean_absolute_error: 2.6650\n",
            "Epoch 128/400\n",
            "5/5 [==============================] - 0s 12ms/step - loss: 14.3303 - mean_absolute_error: 2.9918 - val_loss: 18.7648 - val_mean_absolute_error: 2.7035\n",
            "Epoch 129/400\n",
            "5/5 [==============================] - 0s 18ms/step - loss: 14.3682 - mean_absolute_error: 2.6974 - val_loss: 19.1096 - val_mean_absolute_error: 2.7002\n",
            "Epoch 130/400\n",
            "5/5 [==============================] - 0s 17ms/step - loss: 11.3730 - mean_absolute_error: 2.6158 - val_loss: 18.7089 - val_mean_absolute_error: 2.6202\n",
            "Epoch 131/400\n",
            "5/5 [==============================] - 0s 12ms/step - loss: 13.0145 - mean_absolute_error: 2.6246 - val_loss: 18.7173 - val_mean_absolute_error: 2.6623\n",
            "Epoch 132/400\n",
            "5/5 [==============================] - 0s 16ms/step - loss: 12.2167 - mean_absolute_error: 2.6010 - val_loss: 18.7122 - val_mean_absolute_error: 2.7656\n",
            "Epoch 133/400\n",
            "5/5 [==============================] - 0s 15ms/step - loss: 12.1557 - mean_absolute_error: 2.5431 - val_loss: 18.5537 - val_mean_absolute_error: 2.8225\n",
            "Epoch 134/400\n",
            "5/5 [==============================] - 0s 18ms/step - loss: 13.6580 - mean_absolute_error: 2.8061 - val_loss: 18.4720 - val_mean_absolute_error: 2.8186\n",
            "Epoch 135/400\n",
            "5/5 [==============================] - 0s 12ms/step - loss: 10.8055 - mean_absolute_error: 2.7005 - val_loss: 18.5115 - val_mean_absolute_error: 2.7974\n",
            "Epoch 136/400\n",
            "5/5 [==============================] - 0s 15ms/step - loss: 13.2740 - mean_absolute_error: 2.7008 - val_loss: 18.6454 - val_mean_absolute_error: 2.7640\n",
            "Epoch 137/400\n",
            "5/5 [==============================] - 0s 11ms/step - loss: 14.0554 - mean_absolute_error: 2.6978 - val_loss: 18.6001 - val_mean_absolute_error: 2.7061\n",
            "Epoch 138/400\n",
            "5/5 [==============================] - 0s 15ms/step - loss: 13.2327 - mean_absolute_error: 2.7819 - val_loss: 18.8965 - val_mean_absolute_error: 2.6380\n",
            "Epoch 139/400\n",
            "5/5 [==============================] - 0s 18ms/step - loss: 12.2567 - mean_absolute_error: 2.5750 - val_loss: 18.7608 - val_mean_absolute_error: 2.5606\n",
            "Epoch 140/400\n",
            "5/5 [==============================] - 0s 12ms/step - loss: 13.4147 - mean_absolute_error: 2.6310 - val_loss: 18.2180 - val_mean_absolute_error: 2.5326\n",
            "Epoch 141/400\n",
            "5/5 [==============================] - 0s 12ms/step - loss: 15.9970 - mean_absolute_error: 2.8100 - val_loss: 17.8380 - val_mean_absolute_error: 2.5714\n",
            "Epoch 142/400\n",
            "5/5 [==============================] - 0s 12ms/step - loss: 11.6916 - mean_absolute_error: 2.5069 - val_loss: 18.0135 - val_mean_absolute_error: 2.6154\n",
            "Epoch 143/400\n",
            "5/5 [==============================] - 0s 13ms/step - loss: 12.9545 - mean_absolute_error: 2.6656 - val_loss: 18.6354 - val_mean_absolute_error: 2.7124\n",
            "Epoch 144/400\n",
            "5/5 [==============================] - 0s 16ms/step - loss: 11.8156 - mean_absolute_error: 2.5968 - val_loss: 18.4517 - val_mean_absolute_error: 2.7222\n",
            "Epoch 145/400\n",
            "5/5 [==============================] - 0s 11ms/step - loss: 12.6945 - mean_absolute_error: 2.7861 - val_loss: 17.9803 - val_mean_absolute_error: 2.6777\n",
            "Epoch 146/400\n",
            "5/5 [==============================] - 0s 13ms/step - loss: 11.9229 - mean_absolute_error: 2.7072 - val_loss: 17.8578 - val_mean_absolute_error: 2.6935\n",
            "Epoch 147/400\n",
            "5/5 [==============================] - 0s 11ms/step - loss: 14.0309 - mean_absolute_error: 2.8123 - val_loss: 17.8006 - val_mean_absolute_error: 2.7201\n",
            "Epoch 148/400\n",
            "5/5 [==============================] - 0s 15ms/step - loss: 13.3418 - mean_absolute_error: 2.7695 - val_loss: 18.1983 - val_mean_absolute_error: 2.7632\n",
            "Epoch 149/400\n",
            "5/5 [==============================] - 0s 12ms/step - loss: 13.1478 - mean_absolute_error: 2.7696 - val_loss: 18.1507 - val_mean_absolute_error: 2.6908\n",
            "Epoch 150/400\n",
            "5/5 [==============================] - 0s 13ms/step - loss: 12.7609 - mean_absolute_error: 2.7039 - val_loss: 17.9635 - val_mean_absolute_error: 2.6155\n",
            "Epoch 151/400\n",
            "5/5 [==============================] - 0s 17ms/step - loss: 11.0832 - mean_absolute_error: 2.6297 - val_loss: 17.9909 - val_mean_absolute_error: 2.5652\n",
            "Epoch 152/400\n",
            "5/5 [==============================] - 0s 11ms/step - loss: 11.5523 - mean_absolute_error: 2.5856 - val_loss: 18.6052 - val_mean_absolute_error: 2.6001\n",
            "Epoch 153/400\n",
            "5/5 [==============================] - 0s 11ms/step - loss: 12.6173 - mean_absolute_error: 2.7799 - val_loss: 18.9321 - val_mean_absolute_error: 2.6853\n",
            "Epoch 154/400\n",
            "5/5 [==============================] - 0s 13ms/step - loss: 11.5343 - mean_absolute_error: 2.8051 - val_loss: 18.8542 - val_mean_absolute_error: 2.6822\n",
            "Epoch 155/400\n",
            "5/5 [==============================] - 0s 15ms/step - loss: 11.2204 - mean_absolute_error: 2.6859 - val_loss: 18.8684 - val_mean_absolute_error: 2.6904\n",
            "Epoch 156/400\n",
            "5/5 [==============================] - 0s 16ms/step - loss: 12.2607 - mean_absolute_error: 2.7544 - val_loss: 18.0815 - val_mean_absolute_error: 2.6397\n",
            "Epoch 157/400\n",
            "5/5 [==============================] - 0s 12ms/step - loss: 12.0138 - mean_absolute_error: 2.5980 - val_loss: 17.6217 - val_mean_absolute_error: 2.6712\n",
            "Epoch 158/400\n",
            "5/5 [==============================] - 0s 16ms/step - loss: 11.9555 - mean_absolute_error: 2.8503 - val_loss: 17.8396 - val_mean_absolute_error: 2.7050\n",
            "Epoch 159/400\n",
            "5/5 [==============================] - 0s 11ms/step - loss: 12.8232 - mean_absolute_error: 2.7313 - val_loss: 17.7739 - val_mean_absolute_error: 2.6408\n",
            "Epoch 160/400\n",
            "5/5 [==============================] - 0s 17ms/step - loss: 12.0086 - mean_absolute_error: 2.8608 - val_loss: 17.5927 - val_mean_absolute_error: 2.5567\n",
            "Epoch 161/400\n",
            "5/5 [==============================] - 0s 16ms/step - loss: 10.4780 - mean_absolute_error: 2.5748 - val_loss: 18.0709 - val_mean_absolute_error: 2.5391\n",
            "Epoch 162/400\n",
            "5/5 [==============================] - 0s 16ms/step - loss: 12.0518 - mean_absolute_error: 2.6074 - val_loss: 18.1969 - val_mean_absolute_error: 2.5662\n",
            "Epoch 163/400\n",
            "5/5 [==============================] - 0s 16ms/step - loss: 13.0573 - mean_absolute_error: 2.9025 - val_loss: 18.2637 - val_mean_absolute_error: 2.6293\n",
            "Epoch 164/400\n",
            "5/5 [==============================] - 0s 17ms/step - loss: 11.0024 - mean_absolute_error: 2.4800 - val_loss: 18.2193 - val_mean_absolute_error: 2.7295\n",
            "Epoch 165/400\n",
            "5/5 [==============================] - 0s 16ms/step - loss: 10.2791 - mean_absolute_error: 2.5452 - val_loss: 18.6705 - val_mean_absolute_error: 2.7948\n",
            "Epoch 166/400\n",
            "5/5 [==============================] - 0s 12ms/step - loss: 11.9949 - mean_absolute_error: 2.6842 - val_loss: 18.1587 - val_mean_absolute_error: 2.6666\n",
            "Epoch 167/400\n",
            "5/5 [==============================] - 0s 16ms/step - loss: 12.2491 - mean_absolute_error: 2.6520 - val_loss: 18.1638 - val_mean_absolute_error: 2.6353\n",
            "Epoch 168/400\n",
            "5/5 [==============================] - 0s 11ms/step - loss: 13.6550 - mean_absolute_error: 2.8999 - val_loss: 18.3003 - val_mean_absolute_error: 2.6454\n",
            "Epoch 169/400\n",
            "5/5 [==============================] - 0s 19ms/step - loss: 12.2743 - mean_absolute_error: 2.9494 - val_loss: 20.4721 - val_mean_absolute_error: 2.8179\n",
            "Epoch 170/400\n",
            "5/5 [==============================] - 0s 13ms/step - loss: 12.6701 - mean_absolute_error: 2.8193 - val_loss: 20.8942 - val_mean_absolute_error: 2.6955\n",
            "Epoch 171/400\n",
            "5/5 [==============================] - 0s 12ms/step - loss: 12.7377 - mean_absolute_error: 2.6589 - val_loss: 19.6387 - val_mean_absolute_error: 2.5653\n",
            "Epoch 172/400\n",
            "5/5 [==============================] - 0s 13ms/step - loss: 15.3800 - mean_absolute_error: 3.1287 - val_loss: 19.0566 - val_mean_absolute_error: 2.5446\n",
            "Epoch 173/400\n",
            "5/5 [==============================] - 0s 13ms/step - loss: 11.0132 - mean_absolute_error: 2.3805 - val_loss: 18.5494 - val_mean_absolute_error: 2.5718\n",
            "Epoch 174/400\n",
            "5/5 [==============================] - 0s 13ms/step - loss: 14.3866 - mean_absolute_error: 2.9807 - val_loss: 18.0503 - val_mean_absolute_error: 2.6480\n",
            "Epoch 175/400\n",
            "5/5 [==============================] - 0s 16ms/step - loss: 12.8766 - mean_absolute_error: 2.7633 - val_loss: 18.4725 - val_mean_absolute_error: 2.7406\n",
            "Epoch 176/400\n",
            "5/5 [==============================] - 0s 16ms/step - loss: 13.0532 - mean_absolute_error: 2.7715 - val_loss: 18.2314 - val_mean_absolute_error: 2.7145\n",
            "Epoch 177/400\n",
            "5/5 [==============================] - 0s 11ms/step - loss: 11.9303 - mean_absolute_error: 2.5844 - val_loss: 17.6508 - val_mean_absolute_error: 2.6659\n",
            "Epoch 178/400\n",
            "5/5 [==============================] - 0s 16ms/step - loss: 12.4566 - mean_absolute_error: 2.7131 - val_loss: 17.6409 - val_mean_absolute_error: 2.6735\n",
            "Epoch 179/400\n",
            "5/5 [==============================] - 0s 16ms/step - loss: 11.4143 - mean_absolute_error: 2.5493 - val_loss: 17.7589 - val_mean_absolute_error: 2.6800\n",
            "Epoch 180/400\n",
            "5/5 [==============================] - 0s 12ms/step - loss: 12.1480 - mean_absolute_error: 2.8276 - val_loss: 17.7939 - val_mean_absolute_error: 2.6338\n",
            "Epoch 181/400\n",
            "5/5 [==============================] - 0s 15ms/step - loss: 13.8291 - mean_absolute_error: 2.8237 - val_loss: 17.9065 - val_mean_absolute_error: 2.6108\n",
            "Epoch 182/400\n",
            "5/5 [==============================] - 0s 16ms/step - loss: 12.7656 - mean_absolute_error: 2.6542 - val_loss: 18.4387 - val_mean_absolute_error: 2.6199\n",
            "Epoch 183/400\n",
            "5/5 [==============================] - 0s 11ms/step - loss: 10.9694 - mean_absolute_error: 2.6495 - val_loss: 18.1248 - val_mean_absolute_error: 2.5005\n",
            "Epoch 184/400\n",
            "5/5 [==============================] - 0s 19ms/step - loss: 11.1123 - mean_absolute_error: 2.4134 - val_loss: 18.1444 - val_mean_absolute_error: 2.5037\n",
            "Epoch 185/400\n",
            "5/5 [==============================] - 0s 11ms/step - loss: 10.3644 - mean_absolute_error: 2.3356 - val_loss: 18.8645 - val_mean_absolute_error: 2.6876\n",
            "Epoch 186/400\n",
            "5/5 [==============================] - 0s 12ms/step - loss: 10.4046 - mean_absolute_error: 2.5304 - val_loss: 19.7123 - val_mean_absolute_error: 2.8254\n",
            "Epoch 187/400\n",
            "5/5 [==============================] - 0s 12ms/step - loss: 14.7016 - mean_absolute_error: 2.8552 - val_loss: 19.2807 - val_mean_absolute_error: 2.7509\n",
            "Epoch 188/400\n",
            "5/5 [==============================] - 0s 16ms/step - loss: 12.2241 - mean_absolute_error: 2.6288 - val_loss: 18.7814 - val_mean_absolute_error: 2.6332\n",
            "Epoch 189/400\n",
            "5/5 [==============================] - 0s 16ms/step - loss: 10.9804 - mean_absolute_error: 2.3939 - val_loss: 18.5287 - val_mean_absolute_error: 2.5196\n",
            "Epoch 190/400\n",
            "5/5 [==============================] - 0s 13ms/step - loss: 13.0557 - mean_absolute_error: 2.7816 - val_loss: 18.5215 - val_mean_absolute_error: 2.5051\n",
            "Epoch 190: early stopping\n"
          ]
        }
      ],
      "source": [
        "history_drop = model_drop.fit(X_train, y_train, validation_split=0.2, batch_size=BATCH_SIZE, epochs=NUM_EPOCHS, verbose=1, callbacks=early_stop)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "xZxVpk3RH_6a",
        "outputId": "33f0bcb7-aed0-43eb-f3f4-04d175dcbeac"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/400\n",
            "5/5 [==============================] - 1s 58ms/step - loss: 572.1520 - mean_absolute_error: 21.7238 - val_loss: 581.3503 - val_mean_absolute_error: 20.5968\n",
            "Epoch 2/400\n",
            "5/5 [==============================] - 0s 11ms/step - loss: 524.3610 - mean_absolute_error: 20.8147 - val_loss: 533.2994 - val_mean_absolute_error: 19.3881\n",
            "Epoch 3/400\n",
            "5/5 [==============================] - 0s 12ms/step - loss: 473.0482 - mean_absolute_error: 19.2519 - val_loss: 481.5456 - val_mean_absolute_error: 18.0701\n",
            "Epoch 4/400\n",
            "5/5 [==============================] - 0s 13ms/step - loss: 419.7717 - mean_absolute_error: 17.8974 - val_loss: 423.8727 - val_mean_absolute_error: 16.6314\n",
            "Epoch 5/400\n",
            "5/5 [==============================] - 0s 15ms/step - loss: 357.4525 - mean_absolute_error: 16.6031 - val_loss: 359.7822 - val_mean_absolute_error: 14.9858\n",
            "Epoch 6/400\n",
            "5/5 [==============================] - 0s 15ms/step - loss: 292.5481 - mean_absolute_error: 14.4385 - val_loss: 290.4900 - val_mean_absolute_error: 13.0875\n",
            "Epoch 7/400\n",
            "5/5 [==============================] - 0s 16ms/step - loss: 226.1545 - mean_absolute_error: 12.4286 - val_loss: 220.6836 - val_mean_absolute_error: 10.8783\n",
            "Epoch 8/400\n",
            "5/5 [==============================] - 0s 16ms/step - loss: 158.5579 - mean_absolute_error: 10.0669 - val_loss: 156.9722 - val_mean_absolute_error: 8.6387\n",
            "Epoch 9/400\n",
            "5/5 [==============================] - 0s 12ms/step - loss: 106.1218 - mean_absolute_error: 8.1886 - val_loss: 106.8750 - val_mean_absolute_error: 6.4417\n",
            "Epoch 10/400\n",
            "5/5 [==============================] - 0s 12ms/step - loss: 70.0242 - mean_absolute_error: 6.8162 - val_loss: 78.0262 - val_mean_absolute_error: 4.5986\n",
            "Epoch 11/400\n",
            "5/5 [==============================] - 0s 12ms/step - loss: 55.8421 - mean_absolute_error: 5.7523 - val_loss: 64.8732 - val_mean_absolute_error: 4.0933\n",
            "Epoch 12/400\n",
            "5/5 [==============================] - 0s 11ms/step - loss: 52.1542 - mean_absolute_error: 5.2245 - val_loss: 57.0428 - val_mean_absolute_error: 3.7407\n",
            "Epoch 13/400\n",
            "5/5 [==============================] - 0s 12ms/step - loss: 49.3511 - mean_absolute_error: 5.0283 - val_loss: 49.1438 - val_mean_absolute_error: 3.1183\n",
            "Epoch 14/400\n",
            "5/5 [==============================] - 0s 16ms/step - loss: 40.6596 - mean_absolute_error: 4.4273 - val_loss: 43.1836 - val_mean_absolute_error: 2.9721\n",
            "Epoch 15/400\n",
            "5/5 [==============================] - 0s 18ms/step - loss: 34.6596 - mean_absolute_error: 4.1782 - val_loss: 40.4048 - val_mean_absolute_error: 3.2232\n",
            "Epoch 16/400\n",
            "5/5 [==============================] - 0s 13ms/step - loss: 34.6220 - mean_absolute_error: 4.2541 - val_loss: 39.3555 - val_mean_absolute_error: 3.4258\n",
            "Epoch 17/400\n",
            "5/5 [==============================] - 0s 15ms/step - loss: 28.1045 - mean_absolute_error: 3.7915 - val_loss: 37.5323 - val_mean_absolute_error: 3.3822\n",
            "Epoch 18/400\n",
            "5/5 [==============================] - 0s 12ms/step - loss: 26.6667 - mean_absolute_error: 3.9506 - val_loss: 35.9493 - val_mean_absolute_error: 3.2808\n",
            "Epoch 19/400\n",
            "5/5 [==============================] - 0s 11ms/step - loss: 26.7078 - mean_absolute_error: 3.4874 - val_loss: 34.6614 - val_mean_absolute_error: 3.1345\n",
            "Epoch 20/400\n",
            "5/5 [==============================] - 0s 12ms/step - loss: 26.3458 - mean_absolute_error: 3.4495 - val_loss: 33.7355 - val_mean_absolute_error: 3.0675\n",
            "Epoch 21/400\n",
            "5/5 [==============================] - 0s 11ms/step - loss: 24.1458 - mean_absolute_error: 3.5520 - val_loss: 32.9310 - val_mean_absolute_error: 3.0261\n",
            "Epoch 22/400\n",
            "5/5 [==============================] - 0s 16ms/step - loss: 23.0618 - mean_absolute_error: 3.2793 - val_loss: 32.2144 - val_mean_absolute_error: 3.0456\n",
            "Epoch 23/400\n",
            "5/5 [==============================] - 0s 12ms/step - loss: 27.2251 - mean_absolute_error: 3.5279 - val_loss: 31.4825 - val_mean_absolute_error: 3.0925\n",
            "Epoch 24/400\n",
            "5/5 [==============================] - 0s 15ms/step - loss: 26.3494 - mean_absolute_error: 3.7270 - val_loss: 30.7572 - val_mean_absolute_error: 3.1274\n",
            "Epoch 25/400\n",
            "5/5 [==============================] - 0s 11ms/step - loss: 21.1632 - mean_absolute_error: 3.3901 - val_loss: 29.8531 - val_mean_absolute_error: 3.0715\n",
            "Epoch 26/400\n",
            "5/5 [==============================] - 0s 11ms/step - loss: 25.0148 - mean_absolute_error: 3.4869 - val_loss: 29.2338 - val_mean_absolute_error: 3.0349\n",
            "Epoch 27/400\n",
            "5/5 [==============================] - 0s 12ms/step - loss: 20.5386 - mean_absolute_error: 3.2248 - val_loss: 28.6244 - val_mean_absolute_error: 3.0111\n",
            "Epoch 28/400\n",
            "5/5 [==============================] - 0s 15ms/step - loss: 23.5589 - mean_absolute_error: 3.2198 - val_loss: 28.6614 - val_mean_absolute_error: 3.0929\n",
            "Epoch 29/400\n",
            "5/5 [==============================] - 0s 16ms/step - loss: 24.4106 - mean_absolute_error: 3.8089 - val_loss: 28.2190 - val_mean_absolute_error: 3.1063\n",
            "Epoch 30/400\n",
            "5/5 [==============================] - 0s 17ms/step - loss: 20.7173 - mean_absolute_error: 3.2362 - val_loss: 27.3718 - val_mean_absolute_error: 3.0208\n",
            "Epoch 31/400\n",
            "5/5 [==============================] - 0s 16ms/step - loss: 20.7273 - mean_absolute_error: 3.2032 - val_loss: 27.1613 - val_mean_absolute_error: 3.0353\n",
            "Epoch 32/400\n",
            "5/5 [==============================] - 0s 16ms/step - loss: 18.5405 - mean_absolute_error: 3.1276 - val_loss: 27.0092 - val_mean_absolute_error: 3.0610\n",
            "Epoch 33/400\n",
            "5/5 [==============================] - 0s 15ms/step - loss: 19.8705 - mean_absolute_error: 3.1931 - val_loss: 26.8021 - val_mean_absolute_error: 3.0449\n",
            "Epoch 34/400\n",
            "5/5 [==============================] - 0s 15ms/step - loss: 23.4109 - mean_absolute_error: 3.6326 - val_loss: 26.8735 - val_mean_absolute_error: 3.0539\n",
            "Epoch 35/400\n",
            "5/5 [==============================] - 0s 12ms/step - loss: 21.6479 - mean_absolute_error: 3.1872 - val_loss: 26.5455 - val_mean_absolute_error: 3.0316\n",
            "Epoch 36/400\n",
            "5/5 [==============================] - 0s 11ms/step - loss: 19.3136 - mean_absolute_error: 3.0957 - val_loss: 26.4431 - val_mean_absolute_error: 3.0608\n",
            "Epoch 37/400\n",
            "5/5 [==============================] - 0s 15ms/step - loss: 19.4460 - mean_absolute_error: 3.1689 - val_loss: 26.5109 - val_mean_absolute_error: 3.0887\n",
            "Epoch 38/400\n",
            "5/5 [==============================] - 0s 15ms/step - loss: 20.3016 - mean_absolute_error: 3.2129 - val_loss: 26.2661 - val_mean_absolute_error: 3.0592\n",
            "Epoch 39/400\n",
            "5/5 [==============================] - 0s 14ms/step - loss: 20.4692 - mean_absolute_error: 3.1241 - val_loss: 25.4071 - val_mean_absolute_error: 2.9694\n",
            "Epoch 40/400\n",
            "5/5 [==============================] - 0s 11ms/step - loss: 21.1268 - mean_absolute_error: 3.4129 - val_loss: 24.4675 - val_mean_absolute_error: 2.8765\n",
            "Epoch 41/400\n",
            "5/5 [==============================] - 0s 16ms/step - loss: 18.8817 - mean_absolute_error: 3.0700 - val_loss: 24.0301 - val_mean_absolute_error: 2.8431\n",
            "Epoch 42/400\n",
            "5/5 [==============================] - 0s 15ms/step - loss: 20.3170 - mean_absolute_error: 3.2358 - val_loss: 23.9931 - val_mean_absolute_error: 2.9177\n",
            "Epoch 43/400\n",
            "5/5 [==============================] - 0s 14ms/step - loss: 18.3755 - mean_absolute_error: 2.9686 - val_loss: 24.2684 - val_mean_absolute_error: 2.9990\n",
            "Epoch 44/400\n",
            "5/5 [==============================] - 0s 11ms/step - loss: 19.1618 - mean_absolute_error: 3.1447 - val_loss: 24.4019 - val_mean_absolute_error: 3.0365\n",
            "Epoch 45/400\n",
            "5/5 [==============================] - 0s 17ms/step - loss: 17.3367 - mean_absolute_error: 3.0304 - val_loss: 23.8735 - val_mean_absolute_error: 2.9797\n",
            "Epoch 46/400\n",
            "5/5 [==============================] - 0s 16ms/step - loss: 22.5909 - mean_absolute_error: 3.4913 - val_loss: 23.6583 - val_mean_absolute_error: 2.9820\n",
            "Epoch 47/400\n",
            "5/5 [==============================] - 0s 15ms/step - loss: 18.7734 - mean_absolute_error: 3.0155 - val_loss: 24.1860 - val_mean_absolute_error: 3.0604\n",
            "Epoch 48/400\n",
            "5/5 [==============================] - 0s 19ms/step - loss: 19.4668 - mean_absolute_error: 3.0274 - val_loss: 23.9924 - val_mean_absolute_error: 3.0643\n",
            "Epoch 49/400\n",
            "5/5 [==============================] - 0s 19ms/step - loss: 17.8681 - mean_absolute_error: 2.8828 - val_loss: 23.6327 - val_mean_absolute_error: 3.0410\n",
            "Epoch 50/400\n",
            "5/5 [==============================] - 0s 20ms/step - loss: 20.4445 - mean_absolute_error: 3.1593 - val_loss: 23.2382 - val_mean_absolute_error: 2.9865\n",
            "Epoch 51/400\n",
            "5/5 [==============================] - 0s 19ms/step - loss: 19.5089 - mean_absolute_error: 3.0393 - val_loss: 22.9460 - val_mean_absolute_error: 2.9678\n",
            "Epoch 52/400\n",
            "5/5 [==============================] - 0s 18ms/step - loss: 16.3250 - mean_absolute_error: 3.0198 - val_loss: 22.5388 - val_mean_absolute_error: 2.9510\n",
            "Epoch 53/400\n",
            "5/5 [==============================] - 0s 15ms/step - loss: 21.4647 - mean_absolute_error: 3.3854 - val_loss: 22.0151 - val_mean_absolute_error: 2.8935\n",
            "Epoch 54/400\n",
            "5/5 [==============================] - 0s 13ms/step - loss: 19.9115 - mean_absolute_error: 3.3500 - val_loss: 22.2495 - val_mean_absolute_error: 2.8702\n",
            "Epoch 55/400\n",
            "5/5 [==============================] - 0s 16ms/step - loss: 18.4227 - mean_absolute_error: 3.0554 - val_loss: 22.8218 - val_mean_absolute_error: 2.8465\n",
            "Epoch 56/400\n",
            "5/5 [==============================] - 0s 16ms/step - loss: 19.1736 - mean_absolute_error: 3.0262 - val_loss: 23.1098 - val_mean_absolute_error: 2.8444\n",
            "Epoch 57/400\n",
            "5/5 [==============================] - 0s 17ms/step - loss: 17.3746 - mean_absolute_error: 2.8820 - val_loss: 23.1638 - val_mean_absolute_error: 2.8839\n",
            "Epoch 58/400\n",
            "5/5 [==============================] - 0s 19ms/step - loss: 18.0118 - mean_absolute_error: 2.9904 - val_loss: 22.9386 - val_mean_absolute_error: 2.8974\n",
            "Epoch 59/400\n",
            "5/5 [==============================] - 0s 12ms/step - loss: 18.8576 - mean_absolute_error: 2.8694 - val_loss: 22.8629 - val_mean_absolute_error: 2.9370\n",
            "Epoch 60/400\n",
            "5/5 [==============================] - 0s 16ms/step - loss: 18.1787 - mean_absolute_error: 3.0178 - val_loss: 22.8622 - val_mean_absolute_error: 2.9462\n",
            "Epoch 61/400\n",
            "5/5 [==============================] - 0s 12ms/step - loss: 17.8233 - mean_absolute_error: 3.0663 - val_loss: 22.4078 - val_mean_absolute_error: 2.8857\n",
            "Epoch 62/400\n",
            "5/5 [==============================] - 0s 13ms/step - loss: 17.5621 - mean_absolute_error: 2.9892 - val_loss: 21.3446 - val_mean_absolute_error: 2.7894\n",
            "Epoch 63/400\n",
            "5/5 [==============================] - 0s 12ms/step - loss: 15.8941 - mean_absolute_error: 2.9237 - val_loss: 20.9819 - val_mean_absolute_error: 2.7647\n",
            "Epoch 64/400\n",
            "5/5 [==============================] - 0s 11ms/step - loss: 16.4257 - mean_absolute_error: 2.9342 - val_loss: 21.0876 - val_mean_absolute_error: 2.8069\n",
            "Epoch 65/400\n",
            "5/5 [==============================] - 0s 15ms/step - loss: 16.5581 - mean_absolute_error: 3.0318 - val_loss: 21.5688 - val_mean_absolute_error: 2.8475\n",
            "Epoch 66/400\n",
            "5/5 [==============================] - 0s 15ms/step - loss: 19.3781 - mean_absolute_error: 3.1158 - val_loss: 21.8226 - val_mean_absolute_error: 2.8540\n",
            "Epoch 67/400\n",
            "5/5 [==============================] - 0s 15ms/step - loss: 16.5675 - mean_absolute_error: 2.8967 - val_loss: 21.4678 - val_mean_absolute_error: 2.8445\n",
            "Epoch 68/400\n",
            "5/5 [==============================] - 0s 12ms/step - loss: 18.4029 - mean_absolute_error: 3.0279 - val_loss: 21.2691 - val_mean_absolute_error: 2.8741\n",
            "Epoch 69/400\n",
            "5/5 [==============================] - 0s 16ms/step - loss: 17.2114 - mean_absolute_error: 2.9121 - val_loss: 21.5163 - val_mean_absolute_error: 2.9313\n",
            "Epoch 70/400\n",
            "5/5 [==============================] - 0s 11ms/step - loss: 18.7108 - mean_absolute_error: 2.9413 - val_loss: 21.4745 - val_mean_absolute_error: 2.9286\n",
            "Epoch 71/400\n",
            "5/5 [==============================] - 0s 19ms/step - loss: 15.5829 - mean_absolute_error: 2.9541 - val_loss: 21.0053 - val_mean_absolute_error: 2.8362\n",
            "Epoch 72/400\n",
            "5/5 [==============================] - 0s 19ms/step - loss: 19.6061 - mean_absolute_error: 3.0103 - val_loss: 21.4098 - val_mean_absolute_error: 2.8612\n",
            "Epoch 73/400\n",
            "5/5 [==============================] - 0s 17ms/step - loss: 15.6729 - mean_absolute_error: 2.6394 - val_loss: 21.6903 - val_mean_absolute_error: 2.9164\n",
            "Epoch 74/400\n",
            "5/5 [==============================] - 0s 19ms/step - loss: 17.7979 - mean_absolute_error: 3.2947 - val_loss: 21.6573 - val_mean_absolute_error: 2.9192\n",
            "Epoch 75/400\n",
            "5/5 [==============================] - 0s 20ms/step - loss: 16.7373 - mean_absolute_error: 2.8901 - val_loss: 21.4275 - val_mean_absolute_error: 2.8849\n",
            "Epoch 76/400\n",
            "5/5 [==============================] - 0s 19ms/step - loss: 15.2960 - mean_absolute_error: 3.0680 - val_loss: 21.5629 - val_mean_absolute_error: 2.8732\n",
            "Epoch 77/400\n",
            "5/5 [==============================] - 0s 20ms/step - loss: 15.2042 - mean_absolute_error: 2.9344 - val_loss: 21.5763 - val_mean_absolute_error: 2.8572\n",
            "Epoch 78/400\n",
            "5/5 [==============================] - 0s 19ms/step - loss: 18.5595 - mean_absolute_error: 3.0337 - val_loss: 21.9341 - val_mean_absolute_error: 2.9108\n",
            "Epoch 79/400\n",
            "5/5 [==============================] - 0s 15ms/step - loss: 16.5609 - mean_absolute_error: 2.9159 - val_loss: 21.5806 - val_mean_absolute_error: 2.8427\n",
            "Epoch 80/400\n",
            "5/5 [==============================] - 0s 15ms/step - loss: 16.4414 - mean_absolute_error: 3.0545 - val_loss: 21.5573 - val_mean_absolute_error: 2.8412\n",
            "Epoch 81/400\n",
            "5/5 [==============================] - 0s 16ms/step - loss: 13.2832 - mean_absolute_error: 2.7510 - val_loss: 21.3816 - val_mean_absolute_error: 2.8596\n",
            "Epoch 82/400\n",
            "5/5 [==============================] - 0s 15ms/step - loss: 14.8308 - mean_absolute_error: 2.6919 - val_loss: 20.6371 - val_mean_absolute_error: 2.8128\n",
            "Epoch 83/400\n",
            "5/5 [==============================] - 0s 20ms/step - loss: 16.0809 - mean_absolute_error: 2.7409 - val_loss: 20.0098 - val_mean_absolute_error: 2.7836\n",
            "Epoch 84/400\n",
            "5/5 [==============================] - 0s 20ms/step - loss: 17.2353 - mean_absolute_error: 3.0535 - val_loss: 20.0462 - val_mean_absolute_error: 2.8301\n",
            "Epoch 85/400\n",
            "5/5 [==============================] - 0s 20ms/step - loss: 16.5045 - mean_absolute_error: 2.6954 - val_loss: 20.0376 - val_mean_absolute_error: 2.8138\n",
            "Epoch 86/400\n",
            "5/5 [==============================] - 0s 16ms/step - loss: 17.9547 - mean_absolute_error: 2.9776 - val_loss: 20.0901 - val_mean_absolute_error: 2.8180\n",
            "Epoch 87/400\n",
            "5/5 [==============================] - 0s 20ms/step - loss: 15.4298 - mean_absolute_error: 3.0538 - val_loss: 20.3217 - val_mean_absolute_error: 2.8605\n",
            "Epoch 88/400\n",
            "5/5 [==============================] - 0s 17ms/step - loss: 16.7873 - mean_absolute_error: 2.7893 - val_loss: 20.7726 - val_mean_absolute_error: 2.9421\n",
            "Epoch 89/400\n",
            "5/5 [==============================] - 0s 19ms/step - loss: 17.1358 - mean_absolute_error: 3.0459 - val_loss: 20.8565 - val_mean_absolute_error: 2.9601\n",
            "Epoch 90/400\n",
            "5/5 [==============================] - 0s 21ms/step - loss: 15.9700 - mean_absolute_error: 2.9759 - val_loss: 20.2714 - val_mean_absolute_error: 2.8443\n",
            "Epoch 91/400\n",
            "5/5 [==============================] - 0s 19ms/step - loss: 16.4154 - mean_absolute_error: 2.9640 - val_loss: 19.8832 - val_mean_absolute_error: 2.7961\n",
            "Epoch 92/400\n",
            "5/5 [==============================] - 0s 21ms/step - loss: 14.7428 - mean_absolute_error: 2.7673 - val_loss: 19.8141 - val_mean_absolute_error: 2.7300\n",
            "Epoch 93/400\n",
            "5/5 [==============================] - 0s 16ms/step - loss: 15.0844 - mean_absolute_error: 3.2486 - val_loss: 19.9915 - val_mean_absolute_error: 2.6841\n",
            "Epoch 94/400\n",
            "5/5 [==============================] - 0s 23ms/step - loss: 16.0814 - mean_absolute_error: 2.5906 - val_loss: 20.8045 - val_mean_absolute_error: 2.7232\n",
            "Epoch 95/400\n",
            "5/5 [==============================] - 0s 21ms/step - loss: 15.4585 - mean_absolute_error: 2.8342 - val_loss: 21.1371 - val_mean_absolute_error: 2.7497\n",
            "Epoch 96/400\n",
            "5/5 [==============================] - 0s 20ms/step - loss: 17.0395 - mean_absolute_error: 2.8244 - val_loss: 20.8136 - val_mean_absolute_error: 2.7568\n",
            "Epoch 97/400\n",
            "5/5 [==============================] - 0s 22ms/step - loss: 17.0443 - mean_absolute_error: 2.7208 - val_loss: 20.4321 - val_mean_absolute_error: 2.7415\n",
            "Epoch 98/400\n",
            "5/5 [==============================] - 0s 21ms/step - loss: 16.1875 - mean_absolute_error: 2.8359 - val_loss: 20.1691 - val_mean_absolute_error: 2.7796\n",
            "Epoch 99/400\n",
            "5/5 [==============================] - 0s 18ms/step - loss: 14.3354 - mean_absolute_error: 2.5771 - val_loss: 20.0625 - val_mean_absolute_error: 2.8445\n",
            "Epoch 100/400\n",
            "5/5 [==============================] - 0s 19ms/step - loss: 15.6145 - mean_absolute_error: 2.7820 - val_loss: 20.1577 - val_mean_absolute_error: 2.9302\n",
            "Epoch 101/400\n",
            "5/5 [==============================] - 0s 16ms/step - loss: 16.2671 - mean_absolute_error: 2.8392 - val_loss: 20.3848 - val_mean_absolute_error: 2.9605\n",
            "Epoch 102/400\n",
            "5/5 [==============================] - 0s 12ms/step - loss: 15.4985 - mean_absolute_error: 2.8308 - val_loss: 19.8176 - val_mean_absolute_error: 2.8906\n",
            "Epoch 103/400\n",
            "5/5 [==============================] - 0s 12ms/step - loss: 16.2328 - mean_absolute_error: 2.6047 - val_loss: 19.6290 - val_mean_absolute_error: 2.8878\n",
            "Epoch 104/400\n",
            "5/5 [==============================] - 0s 12ms/step - loss: 16.1883 - mean_absolute_error: 2.7985 - val_loss: 19.3025 - val_mean_absolute_error: 2.8400\n",
            "Epoch 105/400\n",
            "5/5 [==============================] - 0s 11ms/step - loss: 16.1884 - mean_absolute_error: 2.8348 - val_loss: 19.1158 - val_mean_absolute_error: 2.7390\n",
            "Epoch 106/400\n",
            "5/5 [==============================] - 0s 15ms/step - loss: 13.6636 - mean_absolute_error: 2.6398 - val_loss: 19.4488 - val_mean_absolute_error: 2.6919\n",
            "Epoch 107/400\n",
            "5/5 [==============================] - 0s 12ms/step - loss: 14.8971 - mean_absolute_error: 2.7405 - val_loss: 20.1904 - val_mean_absolute_error: 2.7305\n",
            "Epoch 108/400\n",
            "5/5 [==============================] - 0s 11ms/step - loss: 14.1291 - mean_absolute_error: 2.4970 - val_loss: 20.3699 - val_mean_absolute_error: 2.7063\n",
            "Epoch 109/400\n",
            "5/5 [==============================] - 0s 16ms/step - loss: 14.7537 - mean_absolute_error: 2.6689 - val_loss: 20.5778 - val_mean_absolute_error: 2.7206\n",
            "Epoch 110/400\n",
            "5/5 [==============================] - 0s 15ms/step - loss: 14.0875 - mean_absolute_error: 2.4305 - val_loss: 20.0901 - val_mean_absolute_error: 2.6258\n",
            "Epoch 111/400\n",
            "5/5 [==============================] - 0s 16ms/step - loss: 16.0635 - mean_absolute_error: 2.9837 - val_loss: 19.9210 - val_mean_absolute_error: 2.6061\n",
            "Epoch 112/400\n",
            "5/5 [==============================] - 0s 15ms/step - loss: 14.5500 - mean_absolute_error: 2.7212 - val_loss: 19.8218 - val_mean_absolute_error: 2.6543\n",
            "Epoch 113/400\n",
            "5/5 [==============================] - 0s 16ms/step - loss: 14.4210 - mean_absolute_error: 2.7693 - val_loss: 19.6844 - val_mean_absolute_error: 2.7381\n",
            "Epoch 114/400\n",
            "5/5 [==============================] - 0s 14ms/step - loss: 16.0345 - mean_absolute_error: 2.8570 - val_loss: 19.7935 - val_mean_absolute_error: 2.7800\n",
            "Epoch 115/400\n",
            "5/5 [==============================] - 0s 13ms/step - loss: 15.3152 - mean_absolute_error: 2.6669 - val_loss: 20.6375 - val_mean_absolute_error: 2.8437\n",
            "Epoch 116/400\n",
            "5/5 [==============================] - 0s 17ms/step - loss: 16.1103 - mean_absolute_error: 3.0311 - val_loss: 21.1820 - val_mean_absolute_error: 2.8519\n",
            "Epoch 117/400\n",
            "5/5 [==============================] - 0s 13ms/step - loss: 14.0471 - mean_absolute_error: 2.6414 - val_loss: 20.7279 - val_mean_absolute_error: 2.7288\n",
            "Epoch 118/400\n",
            "5/5 [==============================] - 0s 14ms/step - loss: 13.7046 - mean_absolute_error: 2.7494 - val_loss: 20.0184 - val_mean_absolute_error: 2.5966\n",
            "Epoch 119/400\n",
            "5/5 [==============================] - 0s 14ms/step - loss: 18.1397 - mean_absolute_error: 3.0708 - val_loss: 20.4042 - val_mean_absolute_error: 2.6358\n",
            "Epoch 120/400\n",
            "5/5 [==============================] - 0s 13ms/step - loss: 14.2460 - mean_absolute_error: 2.6634 - val_loss: 20.2697 - val_mean_absolute_error: 2.6364\n",
            "Epoch 121/400\n",
            "5/5 [==============================] - 0s 13ms/step - loss: 15.9654 - mean_absolute_error: 2.9230 - val_loss: 20.4674 - val_mean_absolute_error: 2.6321\n",
            "Epoch 122/400\n",
            "5/5 [==============================] - 0s 12ms/step - loss: 13.7435 - mean_absolute_error: 2.7611 - val_loss: 20.5753 - val_mean_absolute_error: 2.5723\n",
            "Epoch 123/400\n",
            "5/5 [==============================] - 0s 14ms/step - loss: 17.0975 - mean_absolute_error: 3.0623 - val_loss: 20.5896 - val_mean_absolute_error: 2.5491\n",
            "Epoch 124/400\n",
            "5/5 [==============================] - 0s 16ms/step - loss: 14.2324 - mean_absolute_error: 2.8117 - val_loss: 20.6131 - val_mean_absolute_error: 2.7154\n",
            "Epoch 125/400\n",
            "5/5 [==============================] - 0s 12ms/step - loss: 14.9268 - mean_absolute_error: 2.7925 - val_loss: 20.3820 - val_mean_absolute_error: 2.8283\n",
            "Epoch 126/400\n",
            "5/5 [==============================] - 0s 13ms/step - loss: 15.6602 - mean_absolute_error: 2.7436 - val_loss: 19.8161 - val_mean_absolute_error: 2.8535\n",
            "Epoch 127/400\n",
            "5/5 [==============================] - 0s 12ms/step - loss: 14.9422 - mean_absolute_error: 2.6115 - val_loss: 19.1903 - val_mean_absolute_error: 2.7991\n",
            "Epoch 128/400\n",
            "5/5 [==============================] - 0s 13ms/step - loss: 14.9384 - mean_absolute_error: 2.7922 - val_loss: 18.9111 - val_mean_absolute_error: 2.7911\n",
            "Epoch 129/400\n",
            "5/5 [==============================] - 0s 12ms/step - loss: 14.4727 - mean_absolute_error: 2.6307 - val_loss: 18.9487 - val_mean_absolute_error: 2.7717\n",
            "Epoch 130/400\n",
            "5/5 [==============================] - 0s 12ms/step - loss: 16.2011 - mean_absolute_error: 2.7678 - val_loss: 19.2205 - val_mean_absolute_error: 2.7393\n",
            "Epoch 131/400\n",
            "5/5 [==============================] - 0s 11ms/step - loss: 14.9311 - mean_absolute_error: 2.7639 - val_loss: 19.8817 - val_mean_absolute_error: 2.7250\n",
            "Epoch 132/400\n",
            "5/5 [==============================] - 0s 13ms/step - loss: 14.4976 - mean_absolute_error: 2.6192 - val_loss: 20.4130 - val_mean_absolute_error: 2.7311\n",
            "Epoch 133/400\n",
            "5/5 [==============================] - 0s 11ms/step - loss: 14.7157 - mean_absolute_error: 2.8201 - val_loss: 20.5914 - val_mean_absolute_error: 2.7210\n",
            "Epoch 134/400\n",
            "5/5 [==============================] - 0s 12ms/step - loss: 15.8138 - mean_absolute_error: 2.9681 - val_loss: 20.5927 - val_mean_absolute_error: 2.6923\n",
            "Epoch 135/400\n",
            "5/5 [==============================] - 0s 15ms/step - loss: 15.4035 - mean_absolute_error: 3.0515 - val_loss: 20.5533 - val_mean_absolute_error: 2.7479\n",
            "Epoch 136/400\n",
            "5/5 [==============================] - 0s 15ms/step - loss: 14.3347 - mean_absolute_error: 2.8132 - val_loss: 20.4952 - val_mean_absolute_error: 2.9650\n",
            "Epoch 137/400\n",
            "5/5 [==============================] - 0s 12ms/step - loss: 14.8101 - mean_absolute_error: 2.8031 - val_loss: 20.9000 - val_mean_absolute_error: 3.0972\n",
            "Epoch 138/400\n",
            "5/5 [==============================] - 0s 15ms/step - loss: 14.6447 - mean_absolute_error: 2.6536 - val_loss: 20.2364 - val_mean_absolute_error: 2.9888\n",
            "Epoch 139/400\n",
            "5/5 [==============================] - 0s 11ms/step - loss: 15.6285 - mean_absolute_error: 2.8334 - val_loss: 19.6382 - val_mean_absolute_error: 2.8293\n",
            "Epoch 140/400\n",
            "5/5 [==============================] - 0s 17ms/step - loss: 13.6930 - mean_absolute_error: 2.5259 - val_loss: 19.6730 - val_mean_absolute_error: 2.8259\n",
            "Epoch 141/400\n",
            "5/5 [==============================] - 0s 15ms/step - loss: 12.9885 - mean_absolute_error: 2.6550 - val_loss: 19.3976 - val_mean_absolute_error: 2.8106\n",
            "Epoch 142/400\n",
            "5/5 [==============================] - 0s 15ms/step - loss: 15.5918 - mean_absolute_error: 2.8736 - val_loss: 19.0938 - val_mean_absolute_error: 2.7980\n",
            "Epoch 143/400\n",
            "5/5 [==============================] - 0s 11ms/step - loss: 13.6268 - mean_absolute_error: 2.5981 - val_loss: 19.0280 - val_mean_absolute_error: 2.8263\n",
            "Epoch 144/400\n",
            "5/5 [==============================] - 0s 11ms/step - loss: 13.4478 - mean_absolute_error: 2.6291 - val_loss: 18.8041 - val_mean_absolute_error: 2.8371\n",
            "Epoch 145/400\n",
            "5/5 [==============================] - 0s 16ms/step - loss: 15.7106 - mean_absolute_error: 2.8753 - val_loss: 18.7141 - val_mean_absolute_error: 2.7811\n",
            "Epoch 146/400\n",
            "5/5 [==============================] - 0s 16ms/step - loss: 13.9400 - mean_absolute_error: 2.6733 - val_loss: 19.1670 - val_mean_absolute_error: 2.8056\n",
            "Epoch 147/400\n",
            "5/5 [==============================] - 0s 12ms/step - loss: 13.3836 - mean_absolute_error: 2.5007 - val_loss: 19.4902 - val_mean_absolute_error: 2.8284\n",
            "Epoch 148/400\n",
            "5/5 [==============================] - 0s 11ms/step - loss: 14.5923 - mean_absolute_error: 2.7471 - val_loss: 19.5735 - val_mean_absolute_error: 2.8231\n",
            "Epoch 149/400\n",
            "5/5 [==============================] - 0s 12ms/step - loss: 13.3504 - mean_absolute_error: 2.5818 - val_loss: 19.8369 - val_mean_absolute_error: 2.8120\n",
            "Epoch 150/400\n",
            "5/5 [==============================] - 0s 11ms/step - loss: 13.1165 - mean_absolute_error: 2.4143 - val_loss: 19.5626 - val_mean_absolute_error: 2.7150\n",
            "Epoch 151/400\n",
            "5/5 [==============================] - 0s 18ms/step - loss: 12.1159 - mean_absolute_error: 2.5816 - val_loss: 19.3723 - val_mean_absolute_error: 2.6474\n",
            "Epoch 152/400\n",
            "5/5 [==============================] - 0s 12ms/step - loss: 13.6882 - mean_absolute_error: 2.6154 - val_loss: 19.6328 - val_mean_absolute_error: 2.6413\n",
            "Epoch 153/400\n",
            "5/5 [==============================] - 0s 16ms/step - loss: 14.4132 - mean_absolute_error: 2.6310 - val_loss: 19.8143 - val_mean_absolute_error: 2.7059\n",
            "Epoch 154/400\n",
            "5/5 [==============================] - 0s 12ms/step - loss: 12.6753 - mean_absolute_error: 2.4353 - val_loss: 20.0807 - val_mean_absolute_error: 2.8521\n",
            "Epoch 155/400\n",
            "5/5 [==============================] - 0s 19ms/step - loss: 14.4976 - mean_absolute_error: 2.8234 - val_loss: 20.2725 - val_mean_absolute_error: 2.9057\n",
            "Epoch 156/400\n",
            "5/5 [==============================] - 0s 12ms/step - loss: 13.5025 - mean_absolute_error: 2.6518 - val_loss: 19.7117 - val_mean_absolute_error: 2.8161\n",
            "Epoch 157/400\n",
            "5/5 [==============================] - 0s 15ms/step - loss: 13.8622 - mean_absolute_error: 2.7429 - val_loss: 18.9257 - val_mean_absolute_error: 2.6209\n",
            "Epoch 158/400\n",
            "5/5 [==============================] - 0s 15ms/step - loss: 14.5166 - mean_absolute_error: 2.5147 - val_loss: 18.8645 - val_mean_absolute_error: 2.5681\n",
            "Epoch 159/400\n",
            "5/5 [==============================] - 0s 15ms/step - loss: 13.1067 - mean_absolute_error: 2.7764 - val_loss: 19.1996 - val_mean_absolute_error: 2.7215\n",
            "Epoch 160/400\n",
            "5/5 [==============================] - 0s 12ms/step - loss: 13.6468 - mean_absolute_error: 2.7304 - val_loss: 20.8600 - val_mean_absolute_error: 2.9996\n",
            "Epoch 161/400\n",
            "5/5 [==============================] - 0s 12ms/step - loss: 16.0718 - mean_absolute_error: 2.9782 - val_loss: 19.6137 - val_mean_absolute_error: 2.8634\n",
            "Epoch 162/400\n",
            "5/5 [==============================] - 0s 16ms/step - loss: 13.0705 - mean_absolute_error: 2.5847 - val_loss: 19.0143 - val_mean_absolute_error: 2.6653\n",
            "Epoch 163/400\n",
            "5/5 [==============================] - 0s 15ms/step - loss: 15.8563 - mean_absolute_error: 2.8623 - val_loss: 18.6751 - val_mean_absolute_error: 2.6080\n",
            "Epoch 164/400\n",
            "5/5 [==============================] - 0s 16ms/step - loss: 11.8356 - mean_absolute_error: 2.3304 - val_loss: 19.4569 - val_mean_absolute_error: 2.7152\n",
            "Epoch 165/400\n",
            "5/5 [==============================] - 0s 16ms/step - loss: 13.2242 - mean_absolute_error: 2.6417 - val_loss: 20.3401 - val_mean_absolute_error: 2.7825\n",
            "Epoch 166/400\n",
            "5/5 [==============================] - 0s 15ms/step - loss: 14.0736 - mean_absolute_error: 2.8435 - val_loss: 20.3103 - val_mean_absolute_error: 2.7043\n",
            "Epoch 167/400\n",
            "5/5 [==============================] - 0s 13ms/step - loss: 15.8475 - mean_absolute_error: 2.8867 - val_loss: 21.1960 - val_mean_absolute_error: 2.6755\n",
            "Epoch 168/400\n",
            "5/5 [==============================] - 0s 13ms/step - loss: 14.0611 - mean_absolute_error: 2.6067 - val_loss: 20.9412 - val_mean_absolute_error: 2.6544\n",
            "Epoch 169/400\n",
            "5/5 [==============================] - 0s 13ms/step - loss: 18.7060 - mean_absolute_error: 2.9056 - val_loss: 19.7003 - val_mean_absolute_error: 2.7281\n",
            "Epoch 170/400\n",
            "5/5 [==============================] - 0s 12ms/step - loss: 12.3327 - mean_absolute_error: 2.4371 - val_loss: 19.3503 - val_mean_absolute_error: 2.7764\n",
            "Epoch 171/400\n",
            "5/5 [==============================] - 0s 15ms/step - loss: 14.5472 - mean_absolute_error: 2.8734 - val_loss: 18.9966 - val_mean_absolute_error: 2.7010\n",
            "Epoch 172/400\n",
            "5/5 [==============================] - 0s 11ms/step - loss: 13.3951 - mean_absolute_error: 2.4731 - val_loss: 19.1684 - val_mean_absolute_error: 2.6065\n",
            "Epoch 173/400\n",
            "5/5 [==============================] - 0s 16ms/step - loss: 12.5224 - mean_absolute_error: 2.6174 - val_loss: 19.7860 - val_mean_absolute_error: 2.5566\n",
            "Epoch 174/400\n",
            "5/5 [==============================] - 0s 12ms/step - loss: 14.4081 - mean_absolute_error: 2.7023 - val_loss: 19.7210 - val_mean_absolute_error: 2.4722\n",
            "Epoch 175/400\n",
            "5/5 [==============================] - 0s 15ms/step - loss: 12.5805 - mean_absolute_error: 2.6491 - val_loss: 19.0643 - val_mean_absolute_error: 2.4463\n",
            "Epoch 176/400\n",
            "5/5 [==============================] - 0s 16ms/step - loss: 14.1685 - mean_absolute_error: 2.6413 - val_loss: 18.9174 - val_mean_absolute_error: 2.5122\n",
            "Epoch 177/400\n",
            "5/5 [==============================] - 0s 12ms/step - loss: 12.6210 - mean_absolute_error: 2.5801 - val_loss: 19.0461 - val_mean_absolute_error: 2.6321\n",
            "Epoch 178/400\n",
            "5/5 [==============================] - 0s 16ms/step - loss: 14.1732 - mean_absolute_error: 2.8737 - val_loss: 18.9869 - val_mean_absolute_error: 2.6727\n",
            "Epoch 179/400\n",
            "5/5 [==============================] - 0s 12ms/step - loss: 11.3064 - mean_absolute_error: 2.4533 - val_loss: 18.8249 - val_mean_absolute_error: 2.7332\n",
            "Epoch 180/400\n",
            "5/5 [==============================] - 0s 13ms/step - loss: 12.6951 - mean_absolute_error: 2.7154 - val_loss: 18.9482 - val_mean_absolute_error: 2.8043\n",
            "Epoch 181/400\n",
            "5/5 [==============================] - 0s 11ms/step - loss: 12.7146 - mean_absolute_error: 2.4778 - val_loss: 19.1703 - val_mean_absolute_error: 2.8611\n",
            "Epoch 182/400\n",
            "5/5 [==============================] - 0s 15ms/step - loss: 14.1528 - mean_absolute_error: 2.5986 - val_loss: 19.4409 - val_mean_absolute_error: 2.8671\n",
            "Epoch 183/400\n",
            "5/5 [==============================] - 0s 17ms/step - loss: 13.0017 - mean_absolute_error: 2.6915 - val_loss: 19.7857 - val_mean_absolute_error: 2.8563\n",
            "Epoch 184/400\n",
            "5/5 [==============================] - 0s 17ms/step - loss: 13.3906 - mean_absolute_error: 2.6432 - val_loss: 20.2484 - val_mean_absolute_error: 2.8292\n",
            "Epoch 185/400\n",
            "5/5 [==============================] - 0s 17ms/step - loss: 11.8412 - mean_absolute_error: 2.5291 - val_loss: 20.9864 - val_mean_absolute_error: 2.9183\n",
            "Epoch 186/400\n",
            "5/5 [==============================] - 0s 16ms/step - loss: 12.2808 - mean_absolute_error: 2.5386 - val_loss: 20.7433 - val_mean_absolute_error: 2.8322\n",
            "Epoch 187/400\n",
            "5/5 [==============================] - 0s 15ms/step - loss: 11.7537 - mean_absolute_error: 2.6939 - val_loss: 20.0123 - val_mean_absolute_error: 2.6811\n",
            "Epoch 188/400\n",
            "5/5 [==============================] - 0s 16ms/step - loss: 14.5390 - mean_absolute_error: 2.6362 - val_loss: 19.5958 - val_mean_absolute_error: 2.6238\n",
            "Epoch 189/400\n",
            "5/5 [==============================] - 0s 12ms/step - loss: 13.8084 - mean_absolute_error: 2.8052 - val_loss: 19.1741 - val_mean_absolute_error: 2.6493\n",
            "Epoch 190/400\n",
            "5/5 [==============================] - 0s 11ms/step - loss: 13.8097 - mean_absolute_error: 2.5589 - val_loss: 19.1573 - val_mean_absolute_error: 2.5806\n",
            "Epoch 191/400\n",
            "5/5 [==============================] - 0s 16ms/step - loss: 12.5595 - mean_absolute_error: 2.4844 - val_loss: 18.7194 - val_mean_absolute_error: 2.4906\n",
            "Epoch 192/400\n",
            "5/5 [==============================] - 0s 15ms/step - loss: 12.5517 - mean_absolute_error: 2.6860 - val_loss: 18.8565 - val_mean_absolute_error: 2.4505\n",
            "Epoch 193/400\n",
            "5/5 [==============================] - 0s 15ms/step - loss: 11.7725 - mean_absolute_error: 2.3230 - val_loss: 19.3018 - val_mean_absolute_error: 2.5083\n",
            "Epoch 193: early stopping\n"
          ]
        }
      ],
      "source": [
        "history_reg_drop = model_reg_drop.fit(X_train, y_train, validation_split=0.2, batch_size=BATCH_SIZE, epochs=NUM_EPOCHS, verbose=1, callbacks=early_stop)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "zVU1HHdmKOuz",
        "outputId": "21b6bde6-ac54-4399-d8cb-c4f2e4d4932e"
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 800x1200 with 8 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ],
      "source": [
        "fig, ax = plt.subplots(4, 2, figsize=(8, 12))\n",
        "\n",
        "ax[0][0].plot(history.epoch, history.history[\"loss\"], label=\"loss\", color=\"red\")\n",
        "ax[0][0].plot(history.epoch, history.history[\"val_loss\"], label=\"val_loss\", color=\"orange\")\n",
        "ax[0][0].set_xlabel(\"epochs\")\n",
        "ax[0][0].set_ylabel(\"loss\")\n",
        "ax[0][0].legend()\n",
        "\n",
        "ax[0][1].plot(history.epoch, history.history[\"mean_absolute_error\"], label=\"mee\", color=\"green\")\n",
        "ax[0][1].plot(history.epoch, history.history[\"val_mean_absolute_error\"], label=\"val_mae\", color=\"blue\")\n",
        "ax[0][1].set_xlabel(\"epochs\")\n",
        "ax[0][1].set_ylabel(\"accuracy\")\n",
        "ax[0][1].legend()\n",
        "\n",
        "ax[1][0].plot(history_reg.epoch, history_reg.history[\"loss\"], label=\"reg_loss\", color=\"red\")\n",
        "ax[1][0].plot(history_reg.epoch, history_reg.history[\"val_loss\"], label=\"reg_val_loss\", color=\"orange\")\n",
        "ax[1][0].set_xlabel(\"epochs\")\n",
        "ax[1][0].set_ylabel(\"loss\")\n",
        "ax[1][0].legend()\n",
        "\n",
        "ax[1][1].plot(history_reg.epoch, history_reg.history[\"mean_absolute_error\"], label=\"reg_mee\", color=\"green\")\n",
        "ax[1][1].plot(history_reg.epoch, history_reg.history[\"val_mean_absolute_error\"], label=\"reg_val_mae\", color=\"blue\")\n",
        "ax[1][1].set_xlabel(\"epochs\")\n",
        "ax[1][1].set_ylabel(\"accuracy\")\n",
        "ax[1][1].legend()\n",
        "\n",
        "ax[2][0].plot(history_drop.epoch, history_drop.history[\"loss\"], label=\"drop_loss\", color=\"red\")\n",
        "ax[2][0].plot(history_drop.epoch, history_drop.history[\"val_loss\"], label=\"drop_val_loss\", color=\"orange\")\n",
        "ax[2][0].set_xlabel(\"epochs\")\n",
        "ax[2][0].set_ylabel(\"loss\")\n",
        "ax[2][0].legend()\n",
        "\n",
        "ax[2][1].plot(history_drop.epoch, history_drop.history[\"mean_absolute_error\"], label=\"drop_mee\", color=\"green\")\n",
        "ax[2][1].plot(history_drop.epoch, history_drop.history[\"val_mean_absolute_error\"], label=\"drop_val_mae\", color=\"blue\")\n",
        "ax[2][1].set_xlabel(\"epochs\")\n",
        "ax[2][1].set_ylabel(\"accuracy\")\n",
        "ax[2][1].legend()\n",
        "\n",
        "ax[3][0].plot(history_reg_drop.epoch, history_reg_drop.history[\"loss\"], label=\"reg_drop_loss\", color=\"red\")\n",
        "ax[3][0].plot(history_reg_drop.epoch, history_reg_drop.history[\"val_loss\"], label=\"reg_drop_val_loss\", color=\"orange\")\n",
        "ax[3][0].set_xlabel(\"epochs\")\n",
        "ax[3][0].set_ylabel(\"loss\")\n",
        "ax[3][0].legend()\n",
        "\n",
        "ax[3][1].plot(history_reg_drop.epoch, history_reg_drop.history[\"mean_absolute_error\"], label=\"reg_drop_mee\", color=\"green\")\n",
        "ax[3][1].plot(history_reg_drop.epoch, history_reg_drop.history[\"val_mean_absolute_error\"], label=\"reg_drop_val_mae\", color=\"blue\")\n",
        "ax[3][1].set_xlabel(\"epochs\")\n",
        "ax[3][1].set_ylabel(\"accuracy\")\n",
        "ax[3][1].legend()\n",
        "\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "L-hHyJtDTqAk"
      },
      "source": [
        "#Ocenjivanje mreže"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "BAirH3nBBZaE",
        "outputId": "344e1b92-fbc3-4dad-a6b1-94c50720b447"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "['loss', 'mean_absolute_error']"
            ]
          },
          "metadata": {},
          "execution_count": 19
        }
      ],
      "source": [
        "model.metrics_names"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "W9IIKr1wLHl2",
        "outputId": "be591e98-867e-46a6-9897-6d167ae3347d"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "3/3 [==============================] - 0s 5ms/step - loss: 15.8430 - mean_absolute_error: 2.5745\n",
            "3/3 [==============================] - 0s 5ms/step - loss: 17.8315 - mean_absolute_error: 2.6248\n",
            "3/3 [==============================] - 0s 5ms/step - loss: 15.0162 - mean_absolute_error: 2.5457\n",
            "3/3 [==============================] - 0s 5ms/step - loss: 16.3664 - mean_absolute_error: 2.5291\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "([15.842985153198242, 2.57452654838562],\n",
              " [17.831544876098633, 2.6248228549957275],\n",
              " [15.016212463378906, 2.5457088947296143],\n",
              " [16.36638832092285, 2.529083013534546])"
            ]
          },
          "metadata": {},
          "execution_count": 20
        }
      ],
      "source": [
        "scores = model.evaluate(X_test, y_test, batch_size=BATCH_SIZE)\n",
        "scores_reg = model_reg.evaluate(X_test, y_test, batch_size=BATCH_SIZE)\n",
        "scores_drop = model_drop.evaluate(X_test, y_test, batch_size=BATCH_SIZE)\n",
        "scores_reg_drop = model_reg_drop.evaluate(X_test, y_test, batch_size=BATCH_SIZE)\n",
        "scores, scores_reg, scores_drop, scores_reg_drop"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ehpabqXSSno-",
        "outputId": "09be8663-4fea-4790-b67f-363958308209"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "6/6 [==============================] - 0s 2ms/step\n",
            "6/6 [==============================] - 0s 2ms/step\n",
            "6/6 [==============================] - 0s 2ms/step\n",
            "6/6 [==============================] - 0s 3ms/step\n"
          ]
        }
      ],
      "source": [
        "y_predict = model.predict(X_test)\n",
        "y_predict_reg = model_reg.predict(X_test)\n",
        "y_predict_drop = model_drop.predict(X_test)\n",
        "y_predict_reg_drop = model_reg_drop.predict(X_test)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "otgTkCqdUPmX",
        "outputId": "20ad7da8-2caa-40a6-d62b-ee5a1b51a736"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(0.7907391290536561, 0.7867928534587831, 0.8016594990399822, 0.804529986273344)"
            ]
          },
          "metadata": {},
          "execution_count": 22
        }
      ],
      "source": [
        "r2 = r2_score(y_test, y_predict)\n",
        "r2_reg = r2_score(y_test, y_predict_reg)\n",
        "r2_drop = r2_score(y_test, y_predict_drop)\n",
        "r2_reg_drop = r2_score(y_test, y_predict_reg_drop)\n",
        "r2, r2_reg, r2_drop, r2_reg_drop"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "KymND6nev7l4",
        "outputId": "1ae184c9-7481-4963-860d-fd435cf60a37"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(339, 13)"
            ]
          },
          "metadata": {},
          "execution_count": 23
        }
      ],
      "source": [
        "X_train.shape"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "OdJm9fHUbiFw"
      },
      "outputs": [],
      "source": [
        "X_new = np.random.randn(input_size)\n",
        "X_new = X_new.reshape(1, 13)\n",
        "X_new = scaler.transform(X_new)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "P7YyrA6PvDpL",
        "outputId": "56c0c3ac-e51d-46d3-975f-6d519bd809b2"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "1/1 [==============================] - 0s 22ms/step\n",
            "1/1 [==============================] - 0s 22ms/step\n",
            "1/1 [==============================] - 0s 21ms/step\n",
            "1/1 [==============================] - 0s 25ms/step\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(array([[50.139698]], dtype=float32),\n",
              " array([[48.00457]], dtype=float32),\n",
              " array([[53.69156]], dtype=float32),\n",
              " array([[50.88674]], dtype=float32))"
            ]
          },
          "metadata": {},
          "execution_count": 25
        }
      ],
      "source": [
        "y_new = model.predict(X_new)\n",
        "y_new_reg = model_reg.predict(X_new)\n",
        "y_new_drop = model_drop.predict(X_new)\n",
        "y_new_reg_drop = model_reg_drop.predict(X_new)\n",
        "y_new, y_new_reg, y_new_drop, y_new_reg_drop"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "a96-L_K6saOa"
      },
      "outputs": [],
      "source": []
    }
  ],
  "metadata": {
    "colab": {
      "provenance": [],
      "authorship_tag": "ABX9TyMHmDjg3JOdckVvjGsfrEQX",
      "include_colab_link": true
    },
    "kernelspec": {
      "display_name": "Python 3",
      "name": "python3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "nbformat": 4,
  "nbformat_minor": 0
}