{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "kernelspec": {
      "display_name": "TensorFlow 2.3 on Python 3.6 (CUDA 10.1)",
      "language": "python",
      "name": "python3"
    },
    "language_info": {
      "codemirror_mode": {
        "name": "ipython",
        "version": 3
      },
      "file_extension": ".py",
      "mimetype": "text/x-python",
      "name": "python",
      "nbconvert_exporter": "python",
      "pygments_lexer": "ipython3",
      "version": "3.6.9"
    },
    "colab": {
      "name": "9-3.regression_in_keras.ipynb",
      "provenance": []
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "UpXTEaeLzrnO"
      },
      "source": [
        "# 케라스로 만드는 회귀 예제\n",
        "\n",
        "매사추세츠 보스턴 지역의 주택 가격 예측하기"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "VN3WDCrAzrnW"
      },
      "source": [
        "[![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/rickiepark/dl-illustrated/blob/master/notebooks/9-3.regression_in_keras.ipynb)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "uFI37QNkzrnW"
      },
      "source": [
        "#### 라이브러리를 적재합니다."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "qBzBQb-szrnW"
      },
      "source": [
        "import numpy as np\n",
        "from tensorflow.keras.datasets import boston_housing\n",
        "from tensorflow.keras.models import Sequential\n",
        "from tensorflow.keras.layers import Dense, Dropout \n",
        "from tensorflow.keras.layers import BatchNormalization "
      ],
      "execution_count": 1,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "5AZDseSwzrnX"
      },
      "source": [
        "#### 데이터를 적재합니다."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "DZZvCNg3zrnX",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "d729ec9e-d563-4827-edc0-3b498dba5ceb"
      },
      "source": [
        "(X_train, y_train), (X_valid, y_valid) = boston_housing.load_data()"
      ],
      "execution_count": 2,
      "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"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "tvr-XETbzrnX",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "91e4654e-750f-4795-95e7-d9a10c661006"
      },
      "source": [
        "X_train.shape"
      ],
      "execution_count": 3,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(404, 13)"
            ]
          },
          "metadata": {},
          "execution_count": 3
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "snbtcHYozrnY",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "941a8897-e994-4c96-a255-67e608fcd070"
      },
      "source": [
        "X_valid.shape"
      ],
      "execution_count": 4,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(102, 13)"
            ]
          },
          "metadata": {},
          "execution_count": 4
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "JwaxiBmlzrnY",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "b5d549ec-e365-471c-e2c1-3e44fb9a2604"
      },
      "source": [
        "X_train[0]"
      ],
      "execution_count": 5,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([  1.23247,   0.     ,   8.14   ,   0.     ,   0.538  ,   6.142  ,\n",
              "        91.7    ,   3.9769 ,   4.     , 307.     ,  21.     , 396.9    ,\n",
              "        18.72   ])"
            ]
          },
          "metadata": {},
          "execution_count": 5
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "d95L3aOezrnZ",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "8ffe19f0-3925-4d27-c440-08dede56adf0"
      },
      "source": [
        "y_train[0]"
      ],
      "execution_count": 6,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "15.2"
            ]
          },
          "metadata": {},
          "execution_count": 6
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "e73SbGiKzrnZ"
      },
      "source": [
        "#### 신경망을 만듭니다."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "trObx-ftzrnZ"
      },
      "source": [
        "model = Sequential()\n",
        "\n",
        "model.add(Dense(32, input_dim=13, activation='relu'))\n",
        "model.add(BatchNormalization())\n",
        "\n",
        "model.add(Dense(16, activation='relu'))\n",
        "model.add(BatchNormalization())\n",
        "model.add(Dropout(0.2))\n",
        "\n",
        "model.add(Dense(1, activation='linear'))"
      ],
      "execution_count": 7,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "HabREuUkzrna",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "0a2c817d-8dcc-4dcb-c40e-6bbe9d6517cc"
      },
      "source": [
        "model.summary()"
      ],
      "execution_count": 8,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Model: \"sequential\"\n",
            "_________________________________________________________________\n",
            " Layer (type)                Output Shape              Param #   \n",
            "=================================================================\n",
            " dense (Dense)               (None, 32)                448       \n",
            "                                                                 \n",
            " batch_normalization (BatchN  (None, 32)               128       \n",
            " ormalization)                                                   \n",
            "                                                                 \n",
            " dense_1 (Dense)             (None, 16)                528       \n",
            "                                                                 \n",
            " batch_normalization_1 (Batc  (None, 16)               64        \n",
            " hNormalization)                                                 \n",
            "                                                                 \n",
            " dropout (Dropout)           (None, 16)                0         \n",
            "                                                                 \n",
            " dense_2 (Dense)             (None, 1)                 17        \n",
            "                                                                 \n",
            "=================================================================\n",
            "Total params: 1,185\n",
            "Trainable params: 1,089\n",
            "Non-trainable params: 96\n",
            "_________________________________________________________________\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "6FF2nyQ6zrna"
      },
      "source": [
        "#### 모델을 설정합니다."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Tm5zaQifzrna"
      },
      "source": [
        "model.compile(loss='mean_squared_error', optimizer='adam')"
      ],
      "execution_count": 9,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "CCebUVgXzrna"
      },
      "source": [
        "#### 훈련!"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "LyjT4M-czrnb",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "0dd81f1a-0f44-470b-93ff-f2b51acb598b"
      },
      "source": [
        "model.fit(X_train, y_train, \n",
        "          batch_size=8, epochs=32, verbose=1, \n",
        "          validation_data=(X_valid, y_valid))"
      ],
      "execution_count": 10,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/32\n",
            "51/51 [==============================] - 4s 12ms/step - loss: 579.8355 - val_loss: 443.8278\n",
            "Epoch 2/32\n",
            "51/51 [==============================] - 0s 8ms/step - loss: 560.0992 - val_loss: 483.1794\n",
            "Epoch 3/32\n",
            "51/51 [==============================] - 0s 8ms/step - loss: 543.0685 - val_loss: 469.9540\n",
            "Epoch 4/32\n",
            "51/51 [==============================] - 0s 7ms/step - loss: 521.0295 - val_loss: 465.4912\n",
            "Epoch 5/32\n",
            "51/51 [==============================] - 0s 8ms/step - loss: 501.8672 - val_loss: 455.2636\n",
            "Epoch 6/32\n",
            "51/51 [==============================] - 0s 8ms/step - loss: 473.1227 - val_loss: 435.2798\n",
            "Epoch 7/32\n",
            "51/51 [==============================] - 0s 9ms/step - loss: 446.5820 - val_loss: 424.8003\n",
            "Epoch 8/32\n",
            "51/51 [==============================] - 0s 8ms/step - loss: 413.9762 - val_loss: 409.0033\n",
            "Epoch 9/32\n",
            "51/51 [==============================] - 0s 8ms/step - loss: 379.3524 - val_loss: 358.2756\n",
            "Epoch 10/32\n",
            "51/51 [==============================] - 0s 7ms/step - loss: 339.5614 - val_loss: 315.7754\n",
            "Epoch 11/32\n",
            "51/51 [==============================] - 0s 7ms/step - loss: 293.6695 - val_loss: 254.9101\n",
            "Epoch 12/32\n",
            "51/51 [==============================] - 0s 7ms/step - loss: 258.0252 - val_loss: 261.8799\n",
            "Epoch 13/32\n",
            "51/51 [==============================] - 0s 5ms/step - loss: 209.6821 - val_loss: 217.7905\n",
            "Epoch 14/32\n",
            "51/51 [==============================] - 0s 4ms/step - loss: 179.9783 - val_loss: 179.9889\n",
            "Epoch 15/32\n",
            "51/51 [==============================] - 0s 8ms/step - loss: 149.2293 - val_loss: 143.5271\n",
            "Epoch 16/32\n",
            "51/51 [==============================] - 0s 5ms/step - loss: 123.8661 - val_loss: 248.8328\n",
            "Epoch 17/32\n",
            "51/51 [==============================] - 0s 4ms/step - loss: 93.4710 - val_loss: 162.2863\n",
            "Epoch 18/32\n",
            "51/51 [==============================] - 0s 4ms/step - loss: 79.5144 - val_loss: 154.6616\n",
            "Epoch 19/32\n",
            "51/51 [==============================] - 0s 4ms/step - loss: 63.8505 - val_loss: 90.0440\n",
            "Epoch 20/32\n",
            "51/51 [==============================] - 0s 5ms/step - loss: 58.4107 - val_loss: 84.3149\n",
            "Epoch 21/32\n",
            "51/51 [==============================] - 0s 5ms/step - loss: 59.3938 - val_loss: 91.1221\n",
            "Epoch 22/32\n",
            "51/51 [==============================] - 0s 5ms/step - loss: 49.0404 - val_loss: 31.6145\n",
            "Epoch 23/32\n",
            "51/51 [==============================] - 0s 5ms/step - loss: 46.7913 - val_loss: 42.7062\n",
            "Epoch 24/32\n",
            "51/51 [==============================] - 0s 4ms/step - loss: 45.1807 - val_loss: 39.2975\n",
            "Epoch 25/32\n",
            "51/51 [==============================] - 0s 5ms/step - loss: 41.6249 - val_loss: 46.8431\n",
            "Epoch 26/32\n",
            "51/51 [==============================] - 0s 4ms/step - loss: 49.3840 - val_loss: 31.6878\n",
            "Epoch 27/32\n",
            "51/51 [==============================] - 0s 4ms/step - loss: 43.3416 - val_loss: 46.2134\n",
            "Epoch 28/32\n",
            "51/51 [==============================] - 0s 4ms/step - loss: 41.7307 - val_loss: 27.0634\n",
            "Epoch 29/32\n",
            "51/51 [==============================] - 0s 5ms/step - loss: 43.3982 - val_loss: 35.7756\n",
            "Epoch 30/32\n",
            "51/51 [==============================] - 0s 4ms/step - loss: 41.0184 - val_loss: 26.2336\n",
            "Epoch 31/32\n",
            "51/51 [==============================] - 0s 5ms/step - loss: 38.7970 - val_loss: 54.1452\n",
            "Epoch 32/32\n",
            "51/51 [==============================] - 0s 4ms/step - loss: 42.0810 - val_loss: 27.2458\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<keras.callbacks.History at 0x7ff00004e1f0>"
            ]
          },
          "metadata": {},
          "execution_count": 10
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "rJkI2p1Dzrnb",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "d225c801-4ab2-418b-a6cb-2642094007d4"
      },
      "source": [
        "X_valid[42]"
      ],
      "execution_count": 11,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([  9.32909,   0.     ,  18.1    ,   0.     ,   0.713  ,   6.185  ,\n",
              "        98.7    ,   2.2616 ,  24.     , 666.     ,  20.2    , 396.9    ,\n",
              "        18.13   ])"
            ]
          },
          "metadata": {},
          "execution_count": 11
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "DxL1_WQqzrnc",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "abfbb88e-136b-4593-8812-cdef9547e627"
      },
      "source": [
        "y_valid[42]"
      ],
      "execution_count": 12,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "14.1"
            ]
          },
          "metadata": {},
          "execution_count": 12
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "TCY9v_ejzrnc",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "4d878489-07ac-46d5-b840-a62d8ac7c89d"
      },
      "source": [
        "model.predict(np.reshape(X_valid[42], [1, 13]))"
      ],
      "execution_count": 13,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "1/1 [==============================] - 0s 92ms/step\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([[17.81344]], dtype=float32)"
            ]
          },
          "metadata": {},
          "execution_count": 13
        }
      ]
    }
  ]
}