chapter04_getting-started-with-neural-networks.ipynb

{
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "1csvZ7gIRydn"
      },
      "source": [
        "이 노트북은 [케라스 창시자에게 배우는 딥러닝 2판](https://tensorflow.blog/kerasdl2/)의 예제 코드를 담고 있습니다.\n",
        "\n",
        "<table align=\"left\">\n",
        "    <tr>\n",
        "        <td>\n",
        "            <a href=\"https://colab.research.google.com/github/rickiepark/deep-learning-with-python-2nd/blob/main/chapter04_getting-started-with-neural-networks.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>\n",
        "        </td>\n",
        "    </tr>\n",
        "</table>"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "K-mJrYOYRydu"
      },
      "source": [
        "# 신경망 시작하기: 분류와 회귀"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "oALyyfgXRydv"
      },
      "source": [
        "## 영화 리뷰 분류: 이진 분류 예제"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "rjZERUadRydw"
      },
      "source": [
        "### IMDB 데이터셋"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "9BQhTZrmRydx"
      },
      "source": [
        "**IMDB 데이터셋 로드하기**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 1,
      "metadata": {
        "execution": {
          "iopub.execute_input": "2022-05-21T00:58:24.141645Z",
          "iopub.status.busy": "2022-05-21T00:58:24.140889Z",
          "iopub.status.idle": "2022-05-21T00:58:29.816007Z",
          "shell.execute_reply": "2022-05-21T00:58:29.814955Z"
        },
        "id": "EJIEvVXqRydy",
        "outputId": "23e3c59a-88c2-420d-80bc-021515a206de",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Downloading data from https://storage.googleapis.com/tensorflow/tf-keras-datasets/imdb.npz\n",
            "17464789/17464789 [==============================] - 0s 0us/step\n"
          ]
        }
      ],
      "source": [
        "from tensorflow.keras.datasets import imdb\n",
        "(train_data, train_labels), (test_data, test_labels) = imdb.load_data(\n",
        "    num_words=10000)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 2,
      "metadata": {
        "execution": {
          "iopub.execute_input": "2022-05-21T00:58:29.821048Z",
          "iopub.status.busy": "2022-05-21T00:58:29.820518Z",
          "iopub.status.idle": "2022-05-21T00:58:29.833665Z",
          "shell.execute_reply": "2022-05-21T00:58:29.832695Z"
        },
        "id": "uAk9jmdMRyd0",
        "outputId": "105b599c-695f-4331-b6a4-fc5d0e454030",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[1,\n",
              " 14,\n",
              " 22,\n",
              " 16,\n",
              " 43,\n",
              " 530,\n",
              " 973,\n",
              " 1622,\n",
              " 1385,\n",
              " 65,\n",
              " 458,\n",
              " 4468,\n",
              " 66,\n",
              " 3941,\n",
              " 4,\n",
              " 173,\n",
              " 36,\n",
              " 256,\n",
              " 5,\n",
              " 25,\n",
              " 100,\n",
              " 43,\n",
              " 838,\n",
              " 112,\n",
              " 50,\n",
              " 670,\n",
              " 2,\n",
              " 9,\n",
              " 35,\n",
              " 480,\n",
              " 284,\n",
              " 5,\n",
              " 150,\n",
              " 4,\n",
              " 172,\n",
              " 112,\n",
              " 167,\n",
              " 2,\n",
              " 336,\n",
              " 385,\n",
              " 39,\n",
              " 4,\n",
              " 172,\n",
              " 4536,\n",
              " 1111,\n",
              " 17,\n",
              " 546,\n",
              " 38,\n",
              " 13,\n",
              " 447,\n",
              " 4,\n",
              " 192,\n",
              " 50,\n",
              " 16,\n",
              " 6,\n",
              " 147,\n",
              " 2025,\n",
              " 19,\n",
              " 14,\n",
              " 22,\n",
              " 4,\n",
              " 1920,\n",
              " 4613,\n",
              " 469,\n",
              " 4,\n",
              " 22,\n",
              " 71,\n",
              " 87,\n",
              " 12,\n",
              " 16,\n",
              " 43,\n",
              " 530,\n",
              " 38,\n",
              " 76,\n",
              " 15,\n",
              " 13,\n",
              " 1247,\n",
              " 4,\n",
              " 22,\n",
              " 17,\n",
              " 515,\n",
              " 17,\n",
              " 12,\n",
              " 16,\n",
              " 626,\n",
              " 18,\n",
              " 2,\n",
              " 5,\n",
              " 62,\n",
              " 386,\n",
              " 12,\n",
              " 8,\n",
              " 316,\n",
              " 8,\n",
              " 106,\n",
              " 5,\n",
              " 4,\n",
              " 2223,\n",
              " 5244,\n",
              " 16,\n",
              " 480,\n",
              " 66,\n",
              " 3785,\n",
              " 33,\n",
              " 4,\n",
              " 130,\n",
              " 12,\n",
              " 16,\n",
              " 38,\n",
              " 619,\n",
              " 5,\n",
              " 25,\n",
              " 124,\n",
              " 51,\n",
              " 36,\n",
              " 135,\n",
              " 48,\n",
              " 25,\n",
              " 1415,\n",
              " 33,\n",
              " 6,\n",
              " 22,\n",
              " 12,\n",
              " 215,\n",
              " 28,\n",
              " 77,\n",
              " 52,\n",
              " 5,\n",
              " 14,\n",
              " 407,\n",
              " 16,\n",
              " 82,\n",
              " 2,\n",
              " 8,\n",
              " 4,\n",
              " 107,\n",
              " 117,\n",
              " 5952,\n",
              " 15,\n",
              " 256,\n",
              " 4,\n",
              " 2,\n",
              " 7,\n",
              " 3766,\n",
              " 5,\n",
              " 723,\n",
              " 36,\n",
              " 71,\n",
              " 43,\n",
              " 530,\n",
              " 476,\n",
              " 26,\n",
              " 400,\n",
              " 317,\n",
              " 46,\n",
              " 7,\n",
              " 4,\n",
              " 2,\n",
              " 1029,\n",
              " 13,\n",
              " 104,\n",
              " 88,\n",
              " 4,\n",
              " 381,\n",
              " 15,\n",
              " 297,\n",
              " 98,\n",
              " 32,\n",
              " 2071,\n",
              " 56,\n",
              " 26,\n",
              " 141,\n",
              " 6,\n",
              " 194,\n",
              " 7486,\n",
              " 18,\n",
              " 4,\n",
              " 226,\n",
              " 22,\n",
              " 21,\n",
              " 134,\n",
              " 476,\n",
              " 26,\n",
              " 480,\n",
              " 5,\n",
              " 144,\n",
              " 30,\n",
              " 5535,\n",
              " 18,\n",
              " 51,\n",
              " 36,\n",
              " 28,\n",
              " 224,\n",
              " 92,\n",
              " 25,\n",
              " 104,\n",
              " 4,\n",
              " 226,\n",
              " 65,\n",
              " 16,\n",
              " 38,\n",
              " 1334,\n",
              " 88,\n",
              " 12,\n",
              " 16,\n",
              " 283,\n",
              " 5,\n",
              " 16,\n",
              " 4472,\n",
              " 113,\n",
              " 103,\n",
              " 32,\n",
              " 15,\n",
              " 16,\n",
              " 5345,\n",
              " 19,\n",
              " 178,\n",
              " 32]"
            ]
          },
          "metadata": {},
          "execution_count": 2
        }
      ],
      "source": [
        "train_data[0]"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 3,
      "metadata": {
        "execution": {
          "iopub.execute_input": "2022-05-21T00:58:29.837815Z",
          "iopub.status.busy": "2022-05-21T00:58:29.837516Z",
          "iopub.status.idle": "2022-05-21T00:58:29.842773Z",
          "shell.execute_reply": "2022-05-21T00:58:29.841816Z"
        },
        "id": "-NAW-Oh2Ryd2",
        "outputId": "45484fcb-739c-4a80-864d-a4046f3269c5",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "1"
            ]
          },
          "metadata": {},
          "execution_count": 3
        }
      ],
      "source": [
        "train_labels[0]"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 4,
      "metadata": {
        "execution": {
          "iopub.execute_input": "2022-05-21T00:58:29.846598Z",
          "iopub.status.busy": "2022-05-21T00:58:29.846314Z",
          "iopub.status.idle": "2022-05-21T00:58:29.960078Z",
          "shell.execute_reply": "2022-05-21T00:58:29.959154Z"
        },
        "id": "_OKegKwsRyd3",
        "outputId": "3c57e02a-441a-4496-c4b0-13928875a811",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "9999"
            ]
          },
          "metadata": {},
          "execution_count": 4
        }
      ],
      "source": [
        "max([max(sequence) for sequence in train_data])"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "1syr49QyRyd4"
      },
      "source": [
        "**리뷰를 다시 텍스트로 디코딩하기**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 5,
      "metadata": {
        "execution": {
          "iopub.execute_input": "2022-05-21T00:58:29.964258Z",
          "iopub.status.busy": "2022-05-21T00:58:29.963607Z",
          "iopub.status.idle": "2022-05-21T00:58:30.665703Z",
          "shell.execute_reply": "2022-05-21T00:58:30.664813Z"
        },
        "id": "FmGd2tohRyd5",
        "outputId": "9a41f634-a54b-4017-a0f6-e7b270fb7a52",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Downloading data from https://storage.googleapis.com/tensorflow/tf-keras-datasets/imdb_word_index.json\n",
            "1641221/1641221 [==============================] - 0s 0us/step\n"
          ]
        }
      ],
      "source": [
        "word_index = imdb.get_word_index()\n",
        "reverse_word_index = dict(\n",
        "    [(value, key) for (key, value) in word_index.items()])\n",
        "decoded_review = \" \".join(\n",
        "    [reverse_word_index.get(i - 3, \"?\") for i in train_data[0]])"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "TL-I0vCkRyd5"
      },
      "source": [
        "### 데이터 준비"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "JMw5jhfYRyd6"
      },
      "source": [
        "**정수 시퀀스를 멀티-핫 인코딩으로 인코딩하기**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 6,
      "metadata": {
        "execution": {
          "iopub.execute_input": "2022-05-21T00:58:30.670103Z",
          "iopub.status.busy": "2022-05-21T00:58:30.669714Z",
          "iopub.status.idle": "2022-05-21T00:58:35.758749Z",
          "shell.execute_reply": "2022-05-21T00:58:35.757888Z"
        },
        "id": "J2YPggO4Ryd6"
      },
      "outputs": [],
      "source": [
        "import numpy as np\n",
        "def vectorize_sequences(sequences, dimension=10000):\n",
        "    results = np.zeros((len(sequences), dimension))\n",
        "    for i, sequence in enumerate(sequences):\n",
        "        for j in sequence:\n",
        "            results[i, j] = 1.\n",
        "    return results\n",
        "x_train = vectorize_sequences(train_data)\n",
        "x_test = vectorize_sequences(test_data)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 7,
      "metadata": {
        "execution": {
          "iopub.execute_input": "2022-05-21T00:58:35.763362Z",
          "iopub.status.busy": "2022-05-21T00:58:35.762546Z",
          "iopub.status.idle": "2022-05-21T00:58:35.768031Z",
          "shell.execute_reply": "2022-05-21T00:58:35.767303Z"
        },
        "id": "cRP3GExeRyd7",
        "outputId": "b8f92e28-cf9c-439f-d421-7905113d3a3a",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([0., 1., 1., ..., 0., 0., 0.])"
            ]
          },
          "metadata": {},
          "execution_count": 7
        }
      ],
      "source": [
        "x_train[0]"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 8,
      "metadata": {
        "execution": {
          "iopub.execute_input": "2022-05-21T00:58:35.771822Z",
          "iopub.status.busy": "2022-05-21T00:58:35.771015Z",
          "iopub.status.idle": "2022-05-21T00:58:35.775500Z",
          "shell.execute_reply": "2022-05-21T00:58:35.774750Z"
        },
        "id": "iBw59ZljRyd7"
      },
      "outputs": [],
      "source": [
        "y_train = np.asarray(train_labels).astype(\"float32\")\n",
        "y_test = np.asarray(test_labels).astype(\"float32\")"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "LANAjB0URyd8"
      },
      "source": [
        "### 신경망 모델 만들기"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "qObD7ZSGRyd9"
      },
      "source": [
        "**모델 정의하기**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 9,
      "metadata": {
        "execution": {
          "iopub.execute_input": "2022-05-21T00:58:35.779731Z",
          "iopub.status.busy": "2022-05-21T00:58:35.779026Z",
          "iopub.status.idle": "2022-05-21T00:58:35.798811Z",
          "shell.execute_reply": "2022-05-21T00:58:35.797834Z"
        },
        "id": "iuTuYC4kRyd9"
      },
      "outputs": [],
      "source": [
        "from tensorflow import keras\n",
        "from tensorflow.keras import layers\n",
        "\n",
        "model = keras.Sequential([\n",
        "    layers.Dense(16, activation=\"relu\"),\n",
        "    layers.Dense(16, activation=\"relu\"),\n",
        "    layers.Dense(1, activation=\"sigmoid\")\n",
        "])"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "JoRYVfdjRyd-"
      },
      "source": [
        "**모델 컴파일하기**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 10,
      "metadata": {
        "execution": {
          "iopub.execute_input": "2022-05-21T00:58:35.803010Z",
          "iopub.status.busy": "2022-05-21T00:58:35.802409Z",
          "iopub.status.idle": "2022-05-21T00:58:35.815382Z",
          "shell.execute_reply": "2022-05-21T00:58:35.814494Z"
        },
        "id": "KBQeB8kXRyd-"
      },
      "outputs": [],
      "source": [
        "model.compile(optimizer=\"rmsprop\",\n",
        "              loss=\"binary_crossentropy\",\n",
        "              metrics=[\"accuracy\"])"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Mjevp6k6Ryd_"
      },
      "source": [
        "### 훈련 검증"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "KdUMmM7SRyd_"
      },
      "source": [
        "**검증 세트 준비하기**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 11,
      "metadata": {
        "execution": {
          "iopub.execute_input": "2022-05-21T00:58:35.819667Z",
          "iopub.status.busy": "2022-05-21T00:58:35.819032Z",
          "iopub.status.idle": "2022-05-21T00:58:35.824128Z",
          "shell.execute_reply": "2022-05-21T00:58:35.823100Z"
        },
        "id": "SgqzuOPoRyeA"
      },
      "outputs": [],
      "source": [
        "x_val = x_train[:10000]\n",
        "partial_x_train = x_train[10000:]\n",
        "y_val = y_train[:10000]\n",
        "partial_y_train = y_train[10000:]"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "deqFApPTRyeA"
      },
      "source": [
        "**모델 훈련하기**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 12,
      "metadata": {
        "execution": {
          "iopub.execute_input": "2022-05-21T00:58:35.828086Z",
          "iopub.status.busy": "2022-05-21T00:58:35.827799Z",
          "iopub.status.idle": "2022-05-21T00:58:47.348964Z",
          "shell.execute_reply": "2022-05-21T00:58:47.348136Z"
        },
        "id": "Wsgo70IxRyeA",
        "outputId": "5e630886-e95b-4c3b-a49e-4e400ff870b2",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/20\n",
            "30/30 [==============================] - 3s 67ms/step - loss: 0.5023 - accuracy: 0.7985 - val_loss: 0.3807 - val_accuracy: 0.8749\n",
            "Epoch 2/20\n",
            "30/30 [==============================] - 1s 49ms/step - loss: 0.3047 - accuracy: 0.9028 - val_loss: 0.3052 - val_accuracy: 0.8897\n",
            "Epoch 3/20\n",
            "30/30 [==============================] - 1s 39ms/step - loss: 0.2248 - accuracy: 0.9280 - val_loss: 0.2884 - val_accuracy: 0.8872\n",
            "Epoch 4/20\n",
            "30/30 [==============================] - 1s 49ms/step - loss: 0.1793 - accuracy: 0.9418 - val_loss: 0.2738 - val_accuracy: 0.8897\n",
            "Epoch 5/20\n",
            "30/30 [==============================] - 1s 48ms/step - loss: 0.1492 - accuracy: 0.9519 - val_loss: 0.2869 - val_accuracy: 0.8861\n",
            "Epoch 6/20\n",
            "30/30 [==============================] - 1s 48ms/step - loss: 0.1240 - accuracy: 0.9614 - val_loss: 0.3502 - val_accuracy: 0.8666\n",
            "Epoch 7/20\n",
            "30/30 [==============================] - 1s 44ms/step - loss: 0.1048 - accuracy: 0.9679 - val_loss: 0.3012 - val_accuracy: 0.8839\n",
            "Epoch 8/20\n",
            "30/30 [==============================] - 1s 39ms/step - loss: 0.0876 - accuracy: 0.9746 - val_loss: 0.3231 - val_accuracy: 0.8846\n",
            "Epoch 9/20\n",
            "30/30 [==============================] - 1s 44ms/step - loss: 0.0714 - accuracy: 0.9801 - val_loss: 0.3622 - val_accuracy: 0.8778\n",
            "Epoch 10/20\n",
            "30/30 [==============================] - 1s 47ms/step - loss: 0.0577 - accuracy: 0.9859 - val_loss: 0.3707 - val_accuracy: 0.8805\n",
            "Epoch 11/20\n",
            "30/30 [==============================] - 1s 48ms/step - loss: 0.0495 - accuracy: 0.9879 - val_loss: 0.3941 - val_accuracy: 0.8771\n",
            "Epoch 12/20\n",
            "30/30 [==============================] - 1s 48ms/step - loss: 0.0412 - accuracy: 0.9901 - val_loss: 0.4190 - val_accuracy: 0.8750\n",
            "Epoch 13/20\n",
            "30/30 [==============================] - 1s 48ms/step - loss: 0.0338 - accuracy: 0.9923 - val_loss: 0.4487 - val_accuracy: 0.8715\n",
            "Epoch 14/20\n",
            "30/30 [==============================] - 1s 47ms/step - loss: 0.0274 - accuracy: 0.9941 - val_loss: 0.4769 - val_accuracy: 0.8709\n",
            "Epoch 15/20\n",
            "30/30 [==============================] - 1s 45ms/step - loss: 0.0194 - accuracy: 0.9969 - val_loss: 0.5097 - val_accuracy: 0.8717\n",
            "Epoch 16/20\n",
            "30/30 [==============================] - 1s 40ms/step - loss: 0.0165 - accuracy: 0.9978 - val_loss: 0.5520 - val_accuracy: 0.8655\n",
            "Epoch 17/20\n",
            "30/30 [==============================] - 1s 41ms/step - loss: 0.0132 - accuracy: 0.9985 - val_loss: 0.5771 - val_accuracy: 0.8698\n",
            "Epoch 18/20\n",
            "30/30 [==============================] - 1s 44ms/step - loss: 0.0112 - accuracy: 0.9986 - val_loss: 0.6029 - val_accuracy: 0.8667\n",
            "Epoch 19/20\n",
            "30/30 [==============================] - 1s 47ms/step - loss: 0.0101 - accuracy: 0.9983 - val_loss: 0.6328 - val_accuracy: 0.8661\n",
            "Epoch 20/20\n",
            "30/30 [==============================] - 1s 46ms/step - loss: 0.0049 - accuracy: 0.9999 - val_loss: 0.7076 - val_accuracy: 0.8613\n"
          ]
        }
      ],
      "source": [
        "history = model.fit(partial_x_train,\n",
        "                    partial_y_train,\n",
        "                    epochs=20,\n",
        "                    batch_size=512,\n",
        "                    validation_data=(x_val, y_val))"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 13,
      "metadata": {
        "execution": {
          "iopub.execute_input": "2022-05-21T00:58:47.353728Z",
          "iopub.status.busy": "2022-05-21T00:58:47.352917Z",
          "iopub.status.idle": "2022-05-21T00:58:47.359185Z",
          "shell.execute_reply": "2022-05-21T00:58:47.358228Z"
        },
        "id": "8hVNM3-7RyeB",
        "outputId": "a439a2eb-9585-4398-81a6-8909ffa0ec9c",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "dict_keys(['loss', 'accuracy', 'val_loss', 'val_accuracy'])"
            ]
          },
          "metadata": {},
          "execution_count": 13
        }
      ],
      "source": [
        "history_dict = history.history\n",
        "history_dict.keys()"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "6HHYvTJZRyeB"
      },
      "source": [
        "**훈련과 검증 손실 그리기**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 14,
      "metadata": {
        "execution": {
          "iopub.execute_input": "2022-05-21T00:58:47.363128Z",
          "iopub.status.busy": "2022-05-21T00:58:47.362820Z",
          "iopub.status.idle": "2022-05-21T00:58:48.134762Z",
          "shell.execute_reply": "2022-05-21T00:58:48.133858Z"
        },
        "id": "wVEVb03vRyeC",
        "outputId": "1983dbac-0fc3-45fd-8dd4-843833384baf",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 295
        }
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYIAAAEWCAYAAABrDZDcAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nO3deXhU5dnH8e8tIMiisqnIFrSAYpEtgEqluC9YUIsLpipFQazUinXhlRYpiq2K1lrRinsVi0utRYXiAhaVqkREFARFBI1bI8pWdrjfP54TGEJWkjMzyfw+1zVX5pw5y53J5NzzLOd5zN0REZHMtUeqAxARkdRSIhARyXBKBCIiGU6JQEQkwykRiIhkOCUCEZEMp0QglcrMppnZhZW9bSqZ2TIzOz6G47qZ/SB6/hcz+21Ztt2N8+SY2Yu7G2cJx+1jZnmVfVxJvpqpDkBSz8zWJizWBTYCW6PlS9x9UlmP5e6nxLFtdefuwyrjOGaWBXwK1HL3LdGxJwFl/htK5lEiENy9fsFzM1sGXOzuLxfezsxqFlxcRKT6UNWQFKug6G9m15rZ18BDZtbQzJ43s3wz+z563iJhn1fN7OLo+SAze93Mxkfbfmpmp+zmtm3MbJaZrTGzl81sgpk9VkzcZYnxBjN7Izrei2bWJOH1881suZmtMLNRJbw/Pc3sazOrkbDuDDObHz3vYWb/MbOVZvaVmd1lZnsWc6yHzezGhOWro32+NLPBhbbta2bvmtlqM/vczMYkvDwr+rnSzNaa2ZEF723C/keZ2RwzWxX9PKqs701JzOzQaP+VZrbAzPolvHaqmS2MjvmFmV0VrW8S/X1Wmtl3Zvaamem6lGR6w6U0BwCNgNbAUMJn5qFouRWwHrirhP17AouBJsAtwANmZrux7ePA20BjYAxwfgnnLEuM5wE/B/YD9gQKLkwdgHui4x8Yna8FRXD3t4D/AccWOu7j0fOtwIjo9zkSOA74RQlxE8VwchTPCUBboHD7xP+AC4B9gb7ApWZ2evRa7+jnvu5e393/U+jYjYAXgDuj3+124AUza1zod9jlvSkl5lrAc8CL0X6/BCaZWftokwcI1YwNgB8CM6L1vwbygKbA/sB1gMa9STIlAinNNuB6d9/o7uvdfYW7/93d17n7GmAc8OMS9l/u7ve5+1bgEaAZ4R++zNuaWSugOzDa3Te5++vAlOJOWMYYH3L3j9x9PfAk0DlaPwB43t1nuftG4LfRe1CcvwEDAcysAXBqtA53f8fd33T3Le6+DLi3iDiKcnYU3wfu/j9C4kv8/V519/fdfZu7z4/OV5bjQkgcH7v7o1FcfwMWAT9J2Ka496YkRwD1gT9Ef6MZwPNE7w2wGehgZnu7+/fuPjdhfTOgtbtvdvfXXAOgJZ0SgZQm3903FCyYWV0zuzeqOllNqIrYN7F6pJCvC564+7roaf1ybnsg8F3COoDPiwu4jDF+nfB8XUJMByYeO7oQryjuXIRv/2eaWW3gTGCuuy+P4mgXVXt8HcVxE6F0UJqdYgCWF/r9eprZzKjqaxUwrIzHLTj28kLrlgPNE5aLe29KjdndE5Nm4nF/SkiSy83s32Z2ZLT+VmAJ8KKZLTWzkWX7NaQyKRFIaQp/O/s10B7o6e57s6MqorjqnsrwFdDIzOomrGtZwvYVifGrxGNH52xc3MbuvpBwwTuFnauFIFQxLQLaRnFctzsxEKq3Ej1OKBG1dPd9gL8kHLe0b9NfEqrMErUCvihDXKUdt2Wh+v3tx3X3Oe7en1Bt9CyhpIG7r3H3X7v7QUA/4EozO66CsUg5KRFIeTUg1LmvjOqbr4/7hNE37FxgjJntGX2b/EkJu1QkxqeB08zsR1HD7lhK/z95HPgVIeE8VSiO1cBaMzsEuLSMMTwJDDKzDlEiKhx/A0IJaYOZ9SAkoAL5hKqsg4o59lSgnZmdZ2Y1zewcoAOhGqci3iKUHq4xs1pm1ofwN5oc/c1yzGwfd99MeE+2AZjZaWb2g6gtaBWhXaWkqjiJgRKBlNcdwF7At8CbwL+SdN4cQoPrCuBG4AnC/Q5F2e0Y3X0BcBnh4v4V8D2hMbMkBXX0M9z924T1VxEu0muA+6KYyxLDtOh3mEGoNplRaJNfAGPNbA0wmujbdbTvOkKbyBtRT5wjCh17BXAaodS0ArgGOK1Q3OXm7psIF/5TCO/73cAF7r4o2uR8YFlURTaM8PeE0Bj+MrAW+A9wt7vPrEgsUn6mdhmpiszsCWCRu8deIhGp7lQikCrBzLqb2cFmtkfUvbI/oa5ZRCpIdxZLVXEA8Ayh4TYPuNTd301tSCLVg6qGREQynKqGREQyXJWrGmrSpIlnZWWlOgwRkSrlnXfe+dbdmxb1WpVLBFlZWeTm5qY6DBGRKsXMCt9Rvp2qhkREMpwSgYhIhlMiEBHJcFWujaAomzdvJi8vjw0bNpS+saRUnTp1aNGiBbVq1Up1KCISqRaJIC8vjwYNGpCVlUXxc55Iqrk7K1asIC8vjzZt2qQ6HBGJVIuqoQ0bNtC4cWMlgTRnZjRu3FglN5E0Uy0SAaAkUEXo7ySSfmJNBGZ2spktNrMlRc08ZGZ/NLN50eMjM1sZZzwiIlXRtm1w1VWwbFk8x48tEUTTAk4gjE/eARgYTQy+nbuPcPfO7t4Z+DNhULEqZ8WKFXTu3JnOnTtzwAEH0Lx58+3LmzZtKnHf3NxcLr/88lLPcdRRR1VKrK+++iqnnXZapRxLRJLjppvgttvgpZfiOX6cJYIewBJ3XxpNWjGZMHRwcQYSTfodt0mTICsL9tgj/Jw0qWLHa9y4MfPmzWPevHkMGzaMESNGbF/ec8892bJlS7H7Zmdnc+edd5Z6jtmzZ1csSBGpkmbMgOuvh5wcuPjieM4RZyJozs4TcOex8wTZ25lZa6ANu87EVPD6UDPLNbPc/Pz8CgU1aRIMHQrLl4N7+Dl0aMWTQWGDBg1i2LBh9OzZk2uuuYa3336bI488ki5dunDUUUexePFiYOdv6GPGjGHw4MH06dOHgw46aKcEUb9+/e3b9+nThwEDBnDIIYeQk5NDwQiyU6dO5ZBDDqFbt25cfvnlpX7z/+677zj99NM5/PDDOeKII5g/fz4A//73v7eXaLp06cKaNWv46quv6N27N507d+aHP/whr732WuW+YSKyiy+/hIEDoX17+MtfIK4mtnTpPnou8LS7by3qRXefCEwEyM7OrtC42aNGwbp1O69bty6sz8kpep/dlZeXx+zZs6lRowarV6/mtddeo2bNmrz88stcd911/P3vf99ln0WLFjFz5kzWrFlD+/btufTSS3fpc//uu++yYMECDjzwQHr16sUbb7xBdnY2l1xyCbNmzaJNmzYMHDiw1Piuv/56unTpwrPPPsuMGTO44IILmDdvHuPHj2fChAn06tWLtWvXUqdOHSZOnMhJJ53EqFGj2Lp1K+sKv4kiUqm2bAlJYO1amDkTou+CsYgzEXwBtExYbhGtK8q5hHliY/fZZ+VbXxFnnXUWNWrUAGDVqlVceOGFfPzxx5gZmzdvLnKfvn37Urt2bWrXrs1+++3HN998Q4sWLXbapkePHtvXde7cmWXLllG/fn0OOuig7f3zBw4cyMSJE0uM7/XXX9+ejI499lhWrFjB6tWr6dWrF1deeSU5OTmceeaZtGjRgu7duzN48GA2b97M6aefTufOnSv03ohIyX7zG5g1Cx59FDp0KH37ioizamgO0NbM2pjZnoSL/ZTCG5nZIUBDwsTVsWvVqnzrK6JevXrbn//2t7/lmGOO4YMPPuC5554rti997dq1tz+vUaNGke0LZdmmIkaOHMn999/P+vXr6dWrF4sWLaJ3797MmjWL5s2bM2jQIP76179W6jlFZIfnn4ebbw7V1j/7Wfzniy0RuPsWYDgwHfgQeNLdF5jZWDPrl7DpucBkT9JUaePGQd26O6+rWzesj9OqVato3jw0kTz88MOVfvz27duzdOlSlkX9y5544olS9zn66KOZFDWOvPrqqzRp0oS9996bTz75hI4dO3LttdfSvXt3Fi1axPLly9l///0ZMmQIF198MXPnzq3030FEQhfRCy6ALl3gT39KzjljbSNw96nA1ELrRhdaHhNnDIUVtAOMGhWqg1q1CkmgstsHCrvmmmu48MILufHGG+nbt2+lH3+vvfbi7rvv5uSTT6ZevXp079691H0KGqcPP/xw6tatyyOPPALAHXfcwcyZM9ljjz047LDDOOWUU5g8eTK33nortWrVon79+ioRiMRg40Y4++xw38BTT0GdOsk5b5Wbszg7O9sLT0zz4Ycfcuihh6YoovSxdu1a6tevj7tz2WWX0bZtW0aMGJHqsHahv5dI0X75S7jrLnjmGTjjjMo9tpm94+7ZRb1WbYaYELjvvvvo3Lkzhx12GKtWreKSSy5JdUgiUkZPPBGSwIgRlZ8ESpMu3UelEowYMSItSwAiUrLFi8PNYkceGRqJk00lAhGRFFq3DgYMgNq1Q6kgFVN1qEQgIpJCl10GCxbAtGnQsmXp28dBJQIRkRR58EF4+OFw89hJJ6UuDiUCEZEUeO+9UBo47rgwqFwqKRFUgmOOOYbp06fvtO6OO+7g0ksvLXafPn36UNAN9tRTT2Xlyl2nYhgzZgzjx48v8dzPPvssCxcu3L48evRoXn755fKEXyQNVy0Sn9Wr4ayzoGHDMOBlNBJNyigRVIKBAwcyefLkndZNnjy5TAO/QRg1dN99992tcxdOBGPHjuX444/frWOJSPzc4aKLYOlSmDwZ9t8/1REpEVSKAQMG8MILL2yfhGbZsmV8+eWXHH300Vx66aVkZ2dz2GGHcX0x5b+srCy+/fZbAMaNG0e7du340Y9+tH2oagj3CHTv3p1OnTrx05/+lHXr1jF79mymTJnC1VdfTefOnfnkk08YNGgQTz/9NACvvPIKXbp0oWPHjgwePJiNGzduP9/1119P165d6dixI4sWLSrx99Nw1SKV56674Omnw4gGvXunOpqg2vUauuIKmDevco/ZuTPccUfxrzdq1IgePXowbdo0+vfvz+TJkzn77LMxM8aNG0ejRo3YunUrxx13HPPnz+fwww8v8jjvvPMOkydPZt68eWzZsoWuXbvSrVs3AM4880yGDBkCwG9+8xseeOABfvnLX9KvXz9OO+00BgwYsNOxNmzYwKBBg3jllVdo164dF1xwAffccw9XXHEFAE2aNGHu3LncfffdjB8/nvvvv7/Y30/DVYtUjrfegl//Gk47Da6+OtXR7KASQSVJrB5KrBZ68skn6dq1K126dGHBggU7VeMU9tprr3HGGWdQt25d9t57b/r12zE23wcffMDRRx9Nx44dmTRpEgsWLCgxnsWLF9OmTRvatWsHwIUXXsisWbO2v37mmWcC0K1bt+0D1RXn9ddf5/zzzweKHq76zjvvZOXKldSsWZPu3bvz0EMPMWbMGN5//30aNGhQ4rFFMsWKFWEcoQMPhEceCTMkpotqVyIo6Zt7nPr378+IESOYO3cu69ato1u3bnz66aeMHz+eOXPm0LBhQwYNGlTs8NOlGTRoEM8++yydOnXi4Ycf5tVXX61QvAVDWVdkGOuRI0fSt29fpk6dSq9evZg+ffr24apfeOEFBg0axJVXXskFF1xQoVhFqro1a8Jw0l9/Da+/Do0apTqinaVRTqra6tevzzHHHMPgwYO3lwZWr15NvXr12Gefffjmm2+YNm1aicfo3bs3zz77LOvXr2fNmjU899xz219bs2YNzZo1Y/PmzduHjgZo0KABa9as2eVY7du3Z9myZSxZsgSARx99lB//+Me79btpuGqR3bN+Pdx+Oxx0EPzrX2FY6TIMDJx01a5EkEoDBw7kjDPO2F5F1KlTJ7p06cIhhxxCy5Yt6dWrV4n7d+3alXPOOYdOnTqx33777TSU9A033EDPnj1p2rQpPXv23H7xP/fccxkyZAh33nnn9kZigDp16vDQQw9x1llnsWXLFrp3786wYcN26/fScNUi5bN5Mzz0EIwdC198ASecADfeCD16pDqyomkYakk6/b2kutq2LYwXNHo0LFkCRxwBN90ExxyT6sg0DLWISKzc4bnnwqxi550XZj187jmYPTs9kkBplAhERCpg5kw46ijo1y+MJPr44/Duu6GLqFmqoyubapMIqloVV6bS30mqi7ffDnX/xx4Ln38OEyfCwoUwcGB6dQ0ti1jDNbOTzWyxmS0xs5HFbHO2mS00swVm9vjunKdOnTqsWLFCF5k05+6sWLGCOsmaiFUkBgsWwJlnQs+e4ebV228P7QFDhqRmLoHKEFuvITOrAUwATgDygDlmNsXdFyZs0xb4P6CXu39vZvvtzrlatGhBXl4e+fn5lRG6xKhOnTq0aNEi1WGIlNvSpTBmDDz2GDRoEHoEXXFFeF7Vxdl9tAewxN2XApjZZKA/kHhr7RBggrt/D+Du/92dE9WqVYs2bdpUMFwRkV3Nmwfjx4feQDVrhqEhrrkGGjdOdWSVJ85E0Bz4PGE5D+hZaJt2AGb2BlADGOPu/4oxJhGRUrnDiy+GBPDyy1C/PgwfHpLAgQemOrrKl+obymoCbYE+QAtglpl1dPedBuc3s6HAUIBWrVolO0YRyRCbNoWhocePh/ffDxf9m2+GoUNhN0eKrxLibCz+AkicgbNFtC5RHjDF3Te7+6fAR4TEsBN3n+ju2e6e3bRp09gCFpHMtGoV3HprGAriwgtDieDhh+HTT0M1UHVOAhBvIpgDtDWzNma2J3AuMKXQNs8SSgOYWRNCVdHSGGMSEdnu88/hqqvCpPHXXAOHHBImkZ8/PySEPfdMdYTJEVvVkLtvMbPhwHRC/f+D7r7AzMYCue4+JXrtRDNbCGwFrnb3FXHFJCICoQH4tttCNZA7nHNOmCega9dUR5Ya1WKsIRGR0rjDSy+FKqCCBuAhQ+BXv4LWrVMdXfxKGmso1Y3FIiKx2rgR/vY3+OMfQ5VPs2bwhz/AJZdU/7r/slIiEJFq6b//hXvugbvvDs9/+MMwNPR552VO3X9ZKRGISLXy/vthpsJJk0JpoG9fGDEijAlUVQaBSzYlAhGp8rZtC719/vhHeOWVMAz04MGh/r99+1RHl/6UCESkyvrf/+Cvfw1TQC5eDM2bh/r/IUPSb17gdKZEICJVTl4e3HVXGPr5++/DPMCPPw4DBlTdEUBTSYlARKqMOXNC9c9TT4XqoDPOCPX/Rx2l+v+KUCIQkbT34Yehu+drr8Hee8Pll4dB4DTocOVQIhCRtOUOjzwCl10G9eqF3kA//3lIBlJ5lAhEJC2tXQu/+AU8+miYAP6xx6rnENDpoIrNrCkimeC996Bbt3AvwO9+F4aGUBKIj0oEIpI23OHee8MUkI0ahXsC+vRJdVTVn0oEIpIWVq0Ko4BeemmoCpo3T0kgWZQIRCTl5syBLl3gmWfCjGAvvAD77ZfqqDKHEoGIpIx7uC+gVy/YujV0D73mGthDV6akUhuBiKTEihWhK+hzz8Hpp8MDD2hYiFRR3hWRpHv9dejcGaZPhzvvDFVCSgKpo0QgIkmzbRv8/vehEbh2bZg9G375Sw0PkWqqGhKRpPjmGzj//HBPwDnnhAHjdIdweoi1RGBmJ5vZYjNbYmYji3h9kJnlm9m86HFxHHFMmgRZWaEBKisrLItIcriH/7lOnUJj8MSJYepIJYH0EVuJwMxqABOAE4A8YI6ZTXH3hYU2fcLdh8cVx6RJMHQorFsXlpcvD8sAOTlxnVVEAN5+O0wO8+ab4U7hl16Cjh1THZUUFmeJoAewxN2XuvsmYDLQP8bzFWnUqB1JoMC6dWG9iMTjyy/hwguhZ0/49FN48MGQFJQE0lOciaA58HnCcl60rrCfmtl8M3vazFpWdhCffVa+9SKy+zZsgJtugnbtYPJkuPZa+Oij0E1U9wakr1T/aZ4Dstz9cOAl4JGiNjKzoWaWa2a5+fn55TpBq1blWy8i5ecOf/87HHpoKG2fcAIsXBimjVRbQPqLMxF8ASR+w28RrdvO3Ve4+8Zo8X6gW1EHcveJ7p7t7tlNmzYtVxDjxoWJrBPVrRvWi0jFvfceHHtsmCayfn14+WX4xz/g4INTHZmUVZyJYA7Q1szamNmewLnAlMQNzKxZwmI/4MPKDiInJ/RSaN069FVu3Tosq6FYpGLy82HYMOjaFd5/H+6+G959F447LtWRSXnF1mvI3beY2XBgOlADeNDdF5jZWCDX3acAl5tZP2AL8B0wKI5YcnJ04RepLJs2wYQJYZ6AtWvDDWHXXw8NG6Y6Mtld5u6pjqFcsrOzPTc3N9VhiGSkqVPhyith8WI46SS4/Xbo0CHVUUlZmNk77p5d1GupbiwWkSpg7lw49VTo2zc0DD//PEybpiRQXSgRiEixcnOhX79wM9h//gO33RbaA/r21fhA1YkSgYjs4q23wsW+e/cwUugNN8CyZaFaaM89Ux2dVDYNOici282eDWPHhuGhGzcON4dddpnuBajulAhEhNdfD72AXn4ZmjQJN4L94hfQoEGqI5NkUCIQyWD//ndIADNnhjmCb701TB5fr16qI5NkUiIQyTDu4cL/u9/BrFlwwAGhG+gll+x6F75kBiUCkQzhHqp+xo4NVUHNmsGf/gRDhsBee6U6Okkl9RoSyQAzZsCPfgQnnhiGhf7zn2HpUrj8ciUBUSIQqdbefDOM/XPccWHo9bvvhk8+geHDoU6dVEcn6UKJQKQamj8f+veHI48MN4DdcQd8/HFoCK5dO9XRSbpRG4FINfLxx2EAuMmTQ9//G28MU0XWr5/qyCSdKRGIVAOffx7u/n3wwfCNf+RIuOoqaNQo1ZFJVaBEIFKF5efD738f6v63bQs3gV13XegSKlJWSgQiVdDKlWEAuDvugHXrYNAgGD06TLwkUl5KBCJVyP/+B3fdBTffDN9/D2efHW4MO+SQVEcmVZkSgUgVsH49PPBAmGv766/DyKA33ABduqQ6MqkOlAhE0tiXX4ZpIe+9F1asgB//GP7+dzjqqFRHJtWJEoFIGsrNDfX/TzwBW7fC6aeHbqC9e2tCGKl8SgQiaWLLFnj22ZAA3ngjDAE9fHiYHP6gg1IdnVRnsd5ZbGYnm9liM1tiZiNL2O6nZuZmVuTEyiLV2cqVMH48HHwwnHUWfPVVSAZ5efDHPyoJSPxiKxGYWQ1gAnACkAfMMbMp7r6w0HYNgF8Bb8UVi0g6+ugjuPNOePjh0BuoT5+wfNppUKNGqqOTTBJniaAHsMTdl7r7JmAy0L+I7W4AbgY2xBiLSFpwh1degZ/8BNq3h/vugwED4N13wxwB/fsrCUjyxZkImgOfJyznReu2M7OuQEt3f6GkA5nZUDPLNbPc/Pz8yo9UJGbr1oXun4cfDscfHyaHv/56WL48lAg6d051hJLJUtZYbGZ7ALcDg0rb1t0nAhMBsrOzPd7IRCrPokXwl7/AI4+EtoDDD4eHHoJzz9Uw0JI+4kwEXwAtE5ZbROsKNAB+CLxqoT/cAcAUM+vn7rkxxiUSq82bQ++fe+4J1T21asGZZ4YhoNX9U9JRnIlgDtDWzNoQEsC5wHkFL7r7KqBJwbKZvQpcpSQgVdXnn8PEiXD//eHu39atw53AF10E+++f6uhEilemNgIzqxdV5WBm7cysn5nVKmkfd98CDAemAx8CT7r7AjMba2b9Khp4ec2bB5ddFkZoFKks27bBv/4VGnmzssKFv2tXeP75MBPYddcpCUj6K2uJYBZwtJk1BF4kfNs/B8gpaSd3nwpMLbRudDHb9iljLLtl9uwwVG+rVnDttXGeSTLBt9+Gsf/vvTfM/du0KVxzDQwdCm3apDo6kfIpayIwd19nZhcBd7v7LWY2L87AKtull8Krr8KoUWGclqOPTnVEUtW4hy8U99wDTz0FmzaFOv9x4+CMMzQFpFRdZU4EZnYkoQRwUbSuSvV2Ngt1t+++G3psvPsu7LdfqqOSdLVxY/im/9FHOx5vvgkffBCmgBw6FIYNg8MOS3WkIhVX1kRwBfB/wD+iev6DgJnxhRWPvfcO3+SOOALOPx+mTYM9Yh1kQ9LZtm1hGIfEi/1HH8HixbBs2c7tSfvvD4ceGhqDBw7UHMBSvZh7+brlR43G9d19dTwhlSw7O9tzcyvWsWjiRLjkkjCe+29+U0mBSVpbvz4M3/zhhzsu+B9/HNYXqFcP2rULd/y2a7fj0bYt7Ltv6mIXqQxm9o67FzmeW5lKBGb2ODAM2EpoKN7bzP7k7rdWXpjJM2QI/Pvf4c7OXr3gmGNSHZHEKTcXLrggJIGaNcMgbu3awQkn7HzBb9ZMffwlM5W1aqiDu682sxxgGjASeAeokonALPT2eOcdOO+80LVUXfyqn82b4aabQsnvgAPghRfCxb9WiR2fRTJPWWvIa0X3DZwOTHH3zUCVHuqhfv3QXrBqVUgGW7emOiKpTIsWhd5hY8aEzgHvvw+nnqokIFKUsiaCe4FlQD1glpm1BlLSRlCZOnYME4HPmBG+NUrxHnkk3Ck7bVqqIynZtm3wpz+FuXw//RSefBIeewwaNkx1ZCLpq9yNxdt3NKsZ3T2cVJXRWJzIHQYNgkcfhRdfDCNDys6efz5MlVizZphFa8KE0Niebj77DH7+85DY+/YNQzw3a5bqqETSQ0mNxWUdYmIfM7u9YChoM7uNUDqo8szCHceHHgo5OWGycNnhzTfh7LPDN+xly+Ckk0L/+auvTp/hOtxDiaVjR3j77ZAAnntOSUCkrMpaNfQgsAY4O3qsBh6KK6hkq1cvtBesXRv6iG9JejknPS1aFL5ZH3hgaGg94AD45z/DmE3jx4cEkdj9MhX++98wsuegQdCpE7z3Hlx8sXr/iJRHWRPBwe5+fTTb2FJ3/x1QrWZS7dAhjBs/a1boVprpvvwSTj45VAdNn77jLuyaNeHPf4bbb4dnngldb7/5JjUx/vOf8MMfwtSpcOutYchnze8rUn5lTQTrzexHBQtm1gtI8XfBynf++WHI4JtuCiNKZqpVq+CUU2DFinCRPfjgnV83gxEjQiKYPz/cqf3hh8mN7+c/D+0WzZuHbsBXXaUpHkV2V1nWtYQAABMoSURBVFkTwTBggpktM7NlwF1AGjYXVtyf/xzqmn/2szD8QKbZuDFcYBcuDBf6bt2K3/b008ONeevXh66aM5Mw6MjMmWGWr7/+NdwV/tZboVQgIruvTInA3d9z907A4cDh7t4FODbWyFJkr71Ce8HGjaH/+ebNqY4oebZtC6WiV18N8+iecELp+3TvHhqUDzwQTjwxNNrGYcWKUAo59tgwyucbb4Quv3vuGc/5RDJJuYZcc/fVCWMMXRlDPGmhffswHtEbb2TOWETucMUVIQnedlvoQVVWWVnhverTJzTajh4djldRW7eGLr3nnBMSzR13wPDhYeTYI46o+PFFJOLuu/UAPt/dfSvy6NatmyfLJZe4g/tzzyXtlCnzhz+E3/XXv979Y2za5D54cDjOeee5b9iwe8f59FP30aPdW7UKx2rUyP3yy93fe2/3YxPJdECuF3c9L+6F0h7AZ7u7b0UeyUwE69e7d+4cLkTLlyfttEn38MM7Lt5bt1bsWNu2ud90Uzje0Ue7f/tt2fZbv9798cfdjzsu7GvmfuKJ7k88sfsJRUR2KCkRlFg1ZGZrzGx1EY81wIGxF1dSrE6dUFWybl0YitgsVINMmpTqyCrPtGmhp9Txx8NDD1V8fgYz+L//g7/9LdzcdeSRsGRJ8du/+26o7mnWLIz59Mkn8LvfhZvXpk8P9ypo5i+ReJU4+qi7N6jIwc3sZOBPhNnM7nf3PxR6fRhwGWF467XAUHdfWJFzVra33gqNqJs2heXly8PsVFC+evR09PbbMGBAuBHrmWcqt+H13HOhZcswqfsRR4Q+/716hde+/z4k0wceCCO/1q4dbgq76KJwX4ImCxJJrt0ea6jUA5vVAD4CTgDyCPMYDEy80JvZ3h41PptZP+AX7n5yScet7LGGSpOVFS7+hTVqFPqvZ2UlLZRK9dFH4cK8995hHt64huFesiSM+vnZZzB2bCgB/OMfoVdW164weHAoCWhQOJF4VXhimt3UA1ji7kujICYD/YHticB3nuWsHmk4tPVnnxW9/rvvoE2bMGftaaeFxxFHhDtv093XX4cxg8zCjXNxzsXwgx/Af/4TJne/9tpwwR8yJHz779w5vvOKSNnFWQhvDnyesJwXrduJmV1mZp8AtwCXF3UgMxtaMOBdfn5+LMEWp1WrotcfeGDoZrnffuHn0UeH5zk58PjjIVGko9Wrw13D+fnhruG2beM/Z+PG8NJLISF8+WW4aU9JQCR9pLw21t0nuPvBwLVAkb323X2iu2e7e3bTpk2TGt+4cVC37s7r6taFW26BK68MQx5/+20Y975fv3DBy8mBpk2hd2+4+WZYsKBy+tVX1MaNoS7+gw/C/L3ZRRYS41G7digx1amTvHOKSNnEmQi+AFomLLeI1hVnMmEGtLSSkxNuLmvdOlSltG4dlhMbivfZB846K9yN+9VX4U7b664Lo5mOHBmGQGjTJoza+cwz4S7ZZNu8Odzs9cor8OCDoWpIRATibSyuSWgsPo6QAOYA57n7goRt2rr7x9HznwDXF9eYUSDZjcUV9cUXoQrm+efDRfh//wsJpVOn0EPm2GNDyWHvvSv3vN99FxqBZ88Od/3OmRPGBLrlljCXgIhklpIai2NLBNGJTwXuIHQffdDdx5nZWMKNDVPM7E/A8cBm4HtgeGKiKEpVSwSJNm8OF+QZM8Jj9uxQXVOjRhjcrSAx9OoV5kgoK/fQC6jgoj979o7RQGvWDJPK9OoV7hU49VSN1S+SiVKWCOJQlRNBYRs2hAbUmTNDYnjrrTApTq1a0LNnSArHHLNr3fr69ZCbu/OFv6C6qWHDMBJor17hZ/fuu7ZxiEjmUSKoItauDRf2GTNCcnjnnXAzW5064aJ+yCFh3dy5O0ZFbddux0W/V68wYJ5uyBKRwpQIqqiVK+G113Ykho8/DjdhFVz0jzwy9E4SESlNqm4okwrad1/4yU/CQ0QkLqpEEBHJcEoEIiIZTolARCTDKRGIiGQ4JQIRkQynRCAikuGUCEREMpwSgYhIhlMiEBHJcEoESTBpUpjbeI89ws9Jk1IdkYjIDhpiImaTJsHQobBuXVhevjwsw86T24iIpIpKBDEbNWpHEiiwbl1YLyKSDpQIYvbZZ+VbLyKSbEoEMWvVqnzrRUSSTYkgZuPG7TpDWN26Yb2ISDpQIohZTg5MnAitW4e5glu3DstqKBaRdBFrIjCzk81ssZktMbORRbx+pZktNLP5ZvaKmbWOM55UycmBZcvCtJPLlikJiEh6iS0RmFkNYAJwCtABGGhmHQpt9i6Q7e6HA08Dt8QVj4iIFC3OEkEPYIm7L3X3TcBkoH/iBu4+090LOle+CbSIMR4RESlCnImgOfB5wnJetK44FwHTinrBzIaaWa6Z5ebn51diiCIikhaNxWb2MyAbuLWo1919ortnu3t206ZNkxuciEg1F+cQE18ALROWW0TrdmJmxwOjgB+7+8YY4xERkSLEWSKYA7Q1szZmtidwLjAlcQMz6wLcC/Rz9//GGEuVpkHrRCROsZUI3H2LmQ0HpgM1gAfdfYGZjQVy3X0KoSqoPvCUmQF85u794oqpKtKgdSISN3P3VMdQLtnZ2Z6bm5vqMJImKytc/Atr3TrckyAiUhZm9o67Zxf1Wlo0FkvxNGidiMRNiSDNadA6EYmbEkGa06B1IhI3JYI0p0HrRCRumqqyCsjJ0YVfROKjEkEG0H0IIlISlQiqOd2HICKlUYmgmhs1akcSKLBuXVgvIgJKBNWe7kMQkdIoEVRzug9BREqjRFDN6T4EESmNEkE1p/sQRKQ0SgQZICcnDFC3bVv4Wd4koO6nItWbuo9KidT9VKT6U4lASqTupyLVnxKBlEjdT0WqPyUCKZG6n4pUf0oEUiJ1PxWp/pQIpESV0f1UvY5E0lusicDMTjazxWa2xMxGFvF6bzOba2ZbzGxAnLHI7qtI99OCXkfLl4P7jl5HSgYi6SO2RGBmNYAJwClAB2CgmXUotNlnwCDg8bjikNRSryOR9BfnfQQ9gCXuvhTAzCYD/YGFBRu4+7LotW0xxiEppF5HIukvzqqh5sDnCct50bpyM7OhZpZrZrn5+fmVEpwkh3odiaS/KtFY7O4T3T3b3bObNm2a6nCkHCqj15Eam0XiFWci+AJombDcIlonGaSivY7U2CwSP3P3eA5sVhP4CDiOkADmAOe5+4Iitn0YeN7dny7tuNnZ2Z6bm1vJ0Uq6ysoKF//CWrcOPZhEpGzM7B13zy7qtdhKBO6+BRgOTAc+BJ509wVmNtbM+kWBdTezPOAs4F4z2yVJSGZTY7NI/GJtI3D3qe7ezt0Pdvdx0brR7j4lej7H3Vu4ez13b+zuh8UZj1Q9ldHYrDYGkZJVicZiyVwVbWxWG4NI6ZQIJK1VtLFZN7SJlC62xuK4qLFYymOPPUJJoDCzMGSGSKZISWOxSDpQG4NI6ZQIpFpTG4NI6ZQIpFpLhzYGlSgk3amNQKQEFW1jKChRJCaTunXLP6eDSEWpjUBkN1W0jUG9lqQqUCIQKUFF2xgq685oVS9JnJQIREpQ0TaGyuq1pAZriZMSgUgpKjJVZ2UMw60Ga4mbEoFIjCpaooCKVy+pRCGlUSIQiVlFShSQHg3WKlFUb0oEImku1Q3WKlFUf0oEImku1Q3WKlFUf0oEIlVAKhus06FEoUQSLyUCkWquqpco0iGRpHr/2Ll7lXp069bNRSR5HnvMvW5d93AZDo+6dcP6sjDbed+Ch1nZ9m/duuj9W7dOTvyp3r+yALlezHU15Rf28j6UCESS77HHwoXXLPwsz0WsohfyVCeSVO/vXrH3v0BJiSDWqiEzO9nMFpvZEjMbWcTrtc3siej1t8wsK854RGT3pLKNoqJVUxVt40j1/snotRVbIjCzGsAE4BSgAzDQzDoU2uwi4Ht3/wHwR+DmuOIRkdSoaBtFqhNJqvdPxsCFcZYIegBL3H2pu28CJgP9C23TH3gkev40cJyZWYwxiUgKVKREkepEkur9K2vgwpLEmQiaA58nLOdF64rcxt23AKuAxoUPZGZDzSzXzHLz8/NjCldE0lUqE0mq96+MgQtLE9vENGY2ADjZ3S+Ols8Herr78IRtPoi2yYuWP4m2+ba442piGhHJJJU1uVGqJqb5AmiZsNwiWlfkNmZWE9gHWBFjTCIiVUplDFxYmpqVd6hdzAHamlkbwgX/XOC8QttMAS4E/gMMAGZ4XEUUEZEqKicn3qlNY0sE7r7FzIYD04EawIPuvsDMxhL6s04BHgAeNbMlwHeEZCEiIkkUZ4kAd58KTC20bnTC8w3AWXHGICIiJdNYQyIiGU6JQEQkwykRiIhkuNjuI4iLmeUDy1MdRzGaAMXeA5EGFF/FpHt8kP4xKr6KqUh8rd29aVEvVLlEkM7MLLe4GzbSgeKrmHSPD9I/RsVXMXHFp6ohEZEMp0QgIpLhlAgq18RUB1AKxVcx6R4fpH+Miq9iYolPbQQiIhlOJQIRkQynRCAikuGUCMrJzFqa2UwzW2hmC8zsV0Vs08fMVpnZvOgxuqhjxRjjMjN7Pzr3LpM3WHBnNFf0fDPrmsTY2ie8L/PMbLWZXVFom6S/f2b2oJn9N5ojo2BdIzN7ycw+jn42LGbfC6NtPjazC5MU261mtij6+/3DzPYtZt8SPwsxxzjGzL5I+DueWsy+Jc5tHmN8TyTEtszM5hWzb6zvYXHXlKR+/oqb1V6Poh9AM6Br9LwB8BHQodA2fYDnUxjjMqBJCa+fCkwDDDgCeCtFcdYAvibc6JLS9w/oDXQFPkhYdwswMno+Eri5iP0aAUujnw2j5w2TENuJQM3o+c1FxVaWz0LMMY4BrirDZ+AT4CBgT+C9wv9PccVX6PXbgNGpeA+Lu6Yk8/OnEkE5uftX7j43er4G+JBdp+BMd/2Bv3rwJrCvmTVLQRzHAZ+4e8rvFHf3WYSh0BMlzqn9CHB6EbueBLzk7t+5+/fAS8DJccfm7i96mN4V4E3CxE8pU8z7VxZlmdu8wkqKL5on/Wzgb5V93rIo4ZqStM+fEkEFmFkW0AV4q4iXjzSz98xsmpkdltTAwIEXzewdMxtaxOtlmU86Gc6l+H++VL5/BfZ396+i518D+xexTTq8l4MJJbyilPZZiNvwqPrqwWKqNtLh/Tsa+MbdPy7m9aS9h4WuKUn7/CkR7CYzqw/8HbjC3VcXenkuobqjE/Bn4Nkkh/cjd+8KnAJcZma9k3z+UpnZnkA/4KkiXk71+7cLD+XwtOtrbWajgC3ApGI2SeVn4R7gYKAz8BWh+iUdDaTk0kBS3sOSrilxf/6UCHaDmdUi/MEmufszhV9399XuvjZ6PhWoZWZNkhWfu38R/fwv8A9C8TtRWeaTjtspwFx3/6bwC6l+/xJ8U1BlFv38bxHbpOy9NLNBwGlATnSh2EUZPguxcfdv3H2ru28D7ivm3Cn9LFqYK/1M4InitknGe1jMNSVpnz8lgnKK6hMfAD5099uL2eaAaDvMrAfhfV6RpPjqmVmDgueERsUPCm02Bbgg6j10BLAqoQiaLMV+C0vl+1dIwZzaRD//WcQ204ETzaxhVPVxYrQuVmZ2MnAN0M/d1xWzTVk+C3HGmNjudEYx594+t3lUSjyX8L4ny/HAInfPK+rFZLyHJVxTkvf5i6slvLo+gB8RimjzgXnR41RgGDAs2mY4sIDQA+JN4KgkxndQdN73ohhGResT4zNgAqG3xvtAdpLfw3qEC/s+CetS+v4RktJXwGZCPetFQGPgFeBj4GWgUbRtNnB/wr6DgSXR4+dJim0JoW644DP4l2jbA4GpJX0Wkvj+PRp9vuYTLmrNCscYLZ9K6CnzSVwxFhVftP7hgs9dwrZJfQ9LuKYk7fOnISZERDKcqoZERDKcEoGISIZTIhARyXBKBCIiGU6JQEQkwykRiETMbKvtPDJqpY2EaWZZiSNfiqSTmqkOQCSNrHf3zqkOQiTZVCIQKUU0Hv0t0Zj0b5vZD6L1WWY2IxpU7RUzaxWt39/CHAHvRY+jokPVMLP7ojHnXzSzvaLtL4/Gop9vZpNT9GtKBlMiENlhr0JVQ+ckvLbK3TsCdwF3ROv+DDzi7ocTBn27M1p/J/BvD4PmdSXckQrQFpjg7ocBK4GfRutHAl2i4wyL65cTKY7uLBaJmNlad69fxPplwLHuvjQaHOxrd29sZt8Shk3YHK3/yt2bmFk+0MLdNyYcI4swbnzbaPlaoJa732hm/wLWEkZZfdajAfdEkkUlApGy8WKel8fGhOdb2dFG15cw9lNXYE40IqZI0igRiJTNOQk//xM9n00YLRMgB3gtev4KcCmAmdUws32KO6iZ7QG0dPeZwLXAPsAupRKROOmbh8gOe9nOE5j/y90LupA2NLP5hG/1A6N1vwQeMrOrgXzg59H6XwETzewiwjf/SwkjXxalBvBYlCwMuNPdV1babyRSBmojEClF1EaQ7e7fpjoWkTioakhEJMOpRCAikuFUIhARyXBKBCIiGU6JQEQkwykRiIhkOCUCEZEM9/8NVlKMsttubgAAAABJRU5ErkJggg==\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ],
      "source": [
        "import matplotlib.pyplot as plt\n",
        "history_dict = history.history\n",
        "loss_values = history_dict[\"loss\"]\n",
        "val_loss_values = history_dict[\"val_loss\"]\n",
        "epochs = range(1, len(loss_values) + 1)\n",
        "plt.plot(epochs, loss_values, \"bo\", label=\"Training loss\")\n",
        "plt.plot(epochs, val_loss_values, \"b\", label=\"Validation loss\")\n",
        "plt.title(\"Training and validation loss\")\n",
        "plt.xlabel(\"Epochs\")\n",
        "plt.ylabel(\"Loss\")\n",
        "plt.legend()\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "vyz4nyT0RyeC"
      },
      "source": [
        "**훈련과 검증 정확도 그리기**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 15,
      "metadata": {
        "execution": {
          "iopub.execute_input": "2022-05-21T00:58:48.138805Z",
          "iopub.status.busy": "2022-05-21T00:58:48.138316Z",
          "iopub.status.idle": "2022-05-21T00:58:48.508495Z",
          "shell.execute_reply": "2022-05-21T00:58:48.507640Z"
        },
        "id": "4lhxzsBfRyeD",
        "outputId": "e99c1989-0c91-4449-e0e0-6b5f76bb5a0e",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 295
        }
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ],
      "source": [
        "plt.clf()\n",
        "acc = history_dict[\"accuracy\"]\n",
        "val_acc = history_dict[\"val_accuracy\"]\n",
        "plt.plot(epochs, acc, \"bo\", label=\"Training acc\")\n",
        "plt.plot(epochs, val_acc, \"b\", label=\"Validation acc\")\n",
        "plt.title(\"Training and validation accuracy\")\n",
        "plt.xlabel(\"Epochs\")\n",
        "plt.ylabel(\"Accuracy\")\n",
        "plt.legend()\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "mLMQIrsARyeE"
      },
      "source": [
        "**모델을 처음부터 다시 훈련하기**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 16,
      "metadata": {
        "execution": {
          "iopub.execute_input": "2022-05-21T00:58:48.512410Z",
          "iopub.status.busy": "2022-05-21T00:58:48.512097Z",
          "iopub.status.idle": "2022-05-21T00:58:53.740167Z",
          "shell.execute_reply": "2022-05-21T00:58:53.739318Z"
        },
        "id": "WdbiqqoXRyeE",
        "outputId": "55456e74-ee18-4745-d0b5-4217a80f9a4e",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/4\n",
            "49/49 [==============================] - 2s 28ms/step - loss: 0.5748 - accuracy: 0.6974\n",
            "Epoch 2/4\n",
            "49/49 [==============================] - 1s 28ms/step - loss: 0.4255 - accuracy: 0.8851\n",
            "Epoch 3/4\n",
            "49/49 [==============================] - 2s 33ms/step - loss: 0.3026 - accuracy: 0.9234\n",
            "Epoch 4/4\n",
            "49/49 [==============================] - 2s 33ms/step - loss: 0.2161 - accuracy: 0.9374\n",
            "782/782 [==============================] - 2s 2ms/step - loss: 0.2978 - accuracy: 0.8839\n"
          ]
        }
      ],
      "source": [
        "model = keras.Sequential([\n",
        "    layers.Dense(16, activation=\"relu\"),\n",
        "    layers.Dense(16, activation=\"relu\"),\n",
        "    layers.Dense(1, activation=\"sigmoid\")\n",
        "])\n",
        "model.compile(optimizer=\"rmsprop\",\n",
        "              loss=\"binary_crossentropy\",\n",
        "              metrics=[\"accuracy\"])\n",
        "model.fit(x_train, y_train, epochs=4, batch_size=512)\n",
        "results = model.evaluate(x_test, y_test)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 17,
      "metadata": {
        "execution": {
          "iopub.execute_input": "2022-05-21T00:58:53.744082Z",
          "iopub.status.busy": "2022-05-21T00:58:53.743592Z",
          "iopub.status.idle": "2022-05-21T00:58:53.748662Z",
          "shell.execute_reply": "2022-05-21T00:58:53.747996Z"
        },
        "id": "AutNeyvsRyeF",
        "outputId": "75bade75-da49-431a-e13a-5b5426e21d19",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[0.2978247404098511, 0.8838800191879272]"
            ]
          },
          "metadata": {},
          "execution_count": 17
        }
      ],
      "source": [
        "results"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "9b6hcHKARyeG"
      },
      "source": [
        "### 훈련된 모델로 새로운 데이터에 대해 예측하기"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 18,
      "metadata": {
        "execution": {
          "iopub.execute_input": "2022-05-21T00:58:53.753112Z",
          "iopub.status.busy": "2022-05-21T00:58:53.752141Z",
          "iopub.status.idle": "2022-05-21T00:58:55.786729Z",
          "shell.execute_reply": "2022-05-21T00:58:55.785781Z"
        },
        "id": "453lRW_wRyeH",
        "outputId": "b08e2250-6718-488a-8cbc-5b00a3eddcc5",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "782/782 [==============================] - 2s 3ms/step\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([[0.15315871],\n",
              "       [0.9586573 ],\n",
              "       [0.75667953],\n",
              "       ...,\n",
              "       [0.1147728 ],\n",
              "       [0.07512458],\n",
              "       [0.4015365 ]], dtype=float32)"
            ]
          },
          "metadata": {},
          "execution_count": 18
        }
      ],
      "source": [
        "model.predict(x_test)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "1GvnURGnRyeL"
      },
      "source": [
        "### 추가 실험"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "wx8OWuKGRyeL"
      },
      "source": [
        "### 정리"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "fjJG1lKSRyeL"
      },
      "source": [
        "## 뉴스 기사 분류: 다중 분류 문제"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "jGWt2Xp0RyeM"
      },
      "source": [
        "### 로이터 데이터셋"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "jxiQdKzgRyeM"
      },
      "source": [
        "**로이터 데이터셋 로드하기**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 19,
      "metadata": {
        "execution": {
          "iopub.execute_input": "2022-05-21T00:58:55.791733Z",
          "iopub.status.busy": "2022-05-21T00:58:55.790885Z",
          "iopub.status.idle": "2022-05-21T00:58:57.939572Z",
          "shell.execute_reply": "2022-05-21T00:58:57.938613Z"
        },
        "id": "8qf7A4Y5RyeM",
        "outputId": "ee7414c0-cc27-4463-927d-69e392d2d864",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Downloading data from https://storage.googleapis.com/tensorflow/tf-keras-datasets/reuters.npz\n",
            "2110848/2110848 [==============================] - 0s 0us/step\n"
          ]
        }
      ],
      "source": [
        "from tensorflow.keras.datasets import reuters\n",
        "(train_data, train_labels), (test_data, test_labels) = reuters.load_data(\n",
        "    num_words=10000)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 20,
      "metadata": {
        "execution": {
          "iopub.execute_input": "2022-05-21T00:58:57.945213Z",
          "iopub.status.busy": "2022-05-21T00:58:57.944384Z",
          "iopub.status.idle": "2022-05-21T00:58:57.949630Z",
          "shell.execute_reply": "2022-05-21T00:58:57.948847Z"
        },
        "id": "AQi1eT92RyeN",
        "outputId": "f7f2015f-1c3f-4146-c4f6-8ab7d4d10f41",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "8982"
            ]
          },
          "metadata": {},
          "execution_count": 20
        }
      ],
      "source": [
        "len(train_data)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 21,
      "metadata": {
        "execution": {
          "iopub.execute_input": "2022-05-21T00:58:57.953105Z",
          "iopub.status.busy": "2022-05-21T00:58:57.952847Z",
          "iopub.status.idle": "2022-05-21T00:58:57.957221Z",
          "shell.execute_reply": "2022-05-21T00:58:57.956497Z"
        },
        "id": "HdsaMadiRyeN",
        "outputId": "f7c83547-3cf8-494d-8d59-97d79cc94c9e",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "2246"
            ]
          },
          "metadata": {},
          "execution_count": 21
        }
      ],
      "source": [
        "len(test_data)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 22,
      "metadata": {
        "execution": {
          "iopub.execute_input": "2022-05-21T00:58:57.960663Z",
          "iopub.status.busy": "2022-05-21T00:58:57.960413Z",
          "iopub.status.idle": "2022-05-21T00:58:57.965646Z",
          "shell.execute_reply": "2022-05-21T00:58:57.964973Z"
        },
        "id": "FWjzpVJWRyeO",
        "outputId": "cefeb688-b09b-4563-b67b-2bc7f83c9f8e",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[1,\n",
              " 245,\n",
              " 273,\n",
              " 207,\n",
              " 156,\n",
              " 53,\n",
              " 74,\n",
              " 160,\n",
              " 26,\n",
              " 14,\n",
              " 46,\n",
              " 296,\n",
              " 26,\n",
              " 39,\n",
              " 74,\n",
              " 2979,\n",
              " 3554,\n",
              " 14,\n",
              " 46,\n",
              " 4689,\n",
              " 4329,\n",
              " 86,\n",
              " 61,\n",
              " 3499,\n",
              " 4795,\n",
              " 14,\n",
              " 61,\n",
              " 451,\n",
              " 4329,\n",
              " 17,\n",
              " 12]"
            ]
          },
          "metadata": {},
          "execution_count": 22
        }
      ],
      "source": [
        "train_data[10]"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ou52UiQ2RyeO"
      },
      "source": [
        "**로이터 데이터셋을 텍스트로 디코딩하기**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 23,
      "metadata": {
        "execution": {
          "iopub.execute_input": "2022-05-21T00:58:57.969038Z",
          "iopub.status.busy": "2022-05-21T00:58:57.968769Z",
          "iopub.status.idle": "2022-05-21T00:58:58.606123Z",
          "shell.execute_reply": "2022-05-21T00:58:58.605366Z"
        },
        "id": "UhI3hCQvRyeO",
        "outputId": "f868a0a4-bcff-411b-ea32-18254c866682",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Downloading data from https://storage.googleapis.com/tensorflow/tf-keras-datasets/reuters_word_index.json\n",
            "550378/550378 [==============================] - 0s 0us/step\n"
          ]
        }
      ],
      "source": [
        "word_index = reuters.get_word_index()\n",
        "reverse_word_index = dict([(value, key) for (key, value) in word_index.items()])\n",
        "decoded_newswire = \" \".join([reverse_word_index.get(i - 3, \"?\") for i in\n",
        "    train_data[0]])"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 24,
      "metadata": {
        "execution": {
          "iopub.execute_input": "2022-05-21T00:58:58.610817Z",
          "iopub.status.busy": "2022-05-21T00:58:58.609932Z",
          "iopub.status.idle": "2022-05-21T00:58:58.615565Z",
          "shell.execute_reply": "2022-05-21T00:58:58.614707Z"
        },
        "id": "_dX4yM26RyeO",
        "outputId": "e04f38d1-b6c0-4992-dad6-e4ab6d973ec5",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "3"
            ]
          },
          "metadata": {},
          "execution_count": 24
        }
      ],
      "source": [
        "train_labels[10]"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Qt1koPzGRyeP"
      },
      "source": [
        "### 데이터 준비"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "bwQGYNQORyeP"
      },
      "source": [
        "**데이터 인코딩하기**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 25,
      "metadata": {
        "execution": {
          "iopub.execute_input": "2022-05-21T00:58:58.619093Z",
          "iopub.status.busy": "2022-05-21T00:58:58.618809Z",
          "iopub.status.idle": "2022-05-21T00:58:58.993179Z",
          "shell.execute_reply": "2022-05-21T00:58:58.992334Z"
        },
        "id": "Z3QCG9WCRyeP"
      },
      "outputs": [],
      "source": [
        "x_train = vectorize_sequences(train_data)\n",
        "x_test = vectorize_sequences(test_data)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "fdQHSLwjRyeQ"
      },
      "source": [
        "**레이블 인코딩하기**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 26,
      "metadata": {
        "execution": {
          "iopub.execute_input": "2022-05-21T00:58:58.998170Z",
          "iopub.status.busy": "2022-05-21T00:58:58.997459Z",
          "iopub.status.idle": "2022-05-21T00:58:59.006393Z",
          "shell.execute_reply": "2022-05-21T00:58:59.005391Z"
        },
        "id": "S-E6ROygRyeQ"
      },
      "outputs": [],
      "source": [
        "def to_one_hot(labels, dimension=46):\n",
        "    results = np.zeros((len(labels), dimension))\n",
        "    for i, label in enumerate(labels):\n",
        "        results[i, label] = 1.\n",
        "    return results\n",
        "y_train = to_one_hot(train_labels)\n",
        "y_test = to_one_hot(test_labels)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 27,
      "metadata": {
        "execution": {
          "iopub.execute_input": "2022-05-21T00:58:59.010347Z",
          "iopub.status.busy": "2022-05-21T00:58:59.010045Z",
          "iopub.status.idle": "2022-05-21T00:58:59.015326Z",
          "shell.execute_reply": "2022-05-21T00:58:59.014567Z"
        },
        "id": "I-7VmPp0RyeQ"
      },
      "outputs": [],
      "source": [
        "from tensorflow.keras.utils import to_categorical\n",
        "y_train = to_categorical(train_labels)\n",
        "y_test = to_categorical(test_labels)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "7CVrNaGrRyeQ"
      },
      "source": [
        "### 모델 구성"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "uv-3olCoRyeR"
      },
      "source": [
        "**모델 정의하기**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 28,
      "metadata": {
        "execution": {
          "iopub.execute_input": "2022-05-21T00:58:59.019228Z",
          "iopub.status.busy": "2022-05-21T00:58:59.018698Z",
          "iopub.status.idle": "2022-05-21T00:58:59.027548Z",
          "shell.execute_reply": "2022-05-21T00:58:59.026784Z"
        },
        "id": "sLqB1E6MRyeR"
      },
      "outputs": [],
      "source": [
        "model = keras.Sequential([\n",
        "    layers.Dense(64, activation=\"relu\"),\n",
        "    layers.Dense(64, activation=\"relu\"),\n",
        "    layers.Dense(46, activation=\"softmax\")\n",
        "])"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "guMVZpUNRyeR"
      },
      "source": [
        "**모델 컴파일하기**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 29,
      "metadata": {
        "execution": {
          "iopub.execute_input": "2022-05-21T00:58:59.031190Z",
          "iopub.status.busy": "2022-05-21T00:58:59.030669Z",
          "iopub.status.idle": "2022-05-21T00:58:59.039143Z",
          "shell.execute_reply": "2022-05-21T00:58:59.038300Z"
        },
        "id": "YXWBBKnPRyeY"
      },
      "outputs": [],
      "source": [
        "model.compile(optimizer=\"rmsprop\",\n",
        "              loss=\"categorical_crossentropy\",\n",
        "              metrics=[\"accuracy\"])"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "QwUx1dQ_RyeY"
      },
      "source": [
        "### 훈련 검증"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "cOeH0p6lRyeZ"
      },
      "source": [
        "**검증 세트 준비하기**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 30,
      "metadata": {
        "execution": {
          "iopub.execute_input": "2022-05-21T00:58:59.043039Z",
          "iopub.status.busy": "2022-05-21T00:58:59.042528Z",
          "iopub.status.idle": "2022-05-21T00:58:59.047024Z",
          "shell.execute_reply": "2022-05-21T00:58:59.046046Z"
        },
        "id": "wqLWYJVlRyeZ"
      },
      "outputs": [],
      "source": [
        "x_val = x_train[:1000]\n",
        "partial_x_train = x_train[1000:]\n",
        "y_val = y_train[:1000]\n",
        "partial_y_train = y_train[1000:]"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "tIRPgk9zRyeZ"
      },
      "source": [
        "**모델 훈련하기**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 31,
      "metadata": {
        "execution": {
          "iopub.execute_input": "2022-05-21T00:58:59.050590Z",
          "iopub.status.busy": "2022-05-21T00:58:59.050211Z",
          "iopub.status.idle": "2022-05-21T00:59:10.041884Z",
          "shell.execute_reply": "2022-05-21T00:59:10.040973Z"
        },
        "id": "aIGzzDunRyeZ",
        "outputId": "45a4b580-71ce-485d-b29a-889d6368f175",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/20\n",
            "16/16 [==============================] - 2s 65ms/step - loss: 2.6362 - accuracy: 0.4465 - val_loss: 1.7501 - val_accuracy: 0.6270\n",
            "Epoch 2/20\n",
            "16/16 [==============================] - 1s 52ms/step - loss: 1.4424 - accuracy: 0.6976 - val_loss: 1.3334 - val_accuracy: 0.7040\n",
            "Epoch 3/20\n",
            "16/16 [==============================] - 1s 59ms/step - loss: 1.0642 - accuracy: 0.7669 - val_loss: 1.1360 - val_accuracy: 0.7510\n",
            "Epoch 4/20\n",
            "16/16 [==============================] - 1s 64ms/step - loss: 0.8317 - accuracy: 0.8226 - val_loss: 1.0182 - val_accuracy: 0.7840\n",
            "Epoch 5/20\n",
            "16/16 [==============================] - 1s 59ms/step - loss: 0.6574 - accuracy: 0.8622 - val_loss: 0.9590 - val_accuracy: 0.8070\n",
            "Epoch 6/20\n",
            "16/16 [==============================] - 1s 53ms/step - loss: 0.5272 - accuracy: 0.8923 - val_loss: 0.9090 - val_accuracy: 0.8130\n",
            "Epoch 7/20\n",
            "16/16 [==============================] - 1s 52ms/step - loss: 0.4223 - accuracy: 0.9121 - val_loss: 0.9174 - val_accuracy: 0.8020\n",
            "Epoch 8/20\n",
            "16/16 [==============================] - 1s 52ms/step - loss: 0.3434 - accuracy: 0.9290 - val_loss: 0.8842 - val_accuracy: 0.8140\n",
            "Epoch 9/20\n",
            "16/16 [==============================] - 1s 54ms/step - loss: 0.2824 - accuracy: 0.9385 - val_loss: 0.8904 - val_accuracy: 0.8100\n",
            "Epoch 10/20\n",
            "16/16 [==============================] - 1s 53ms/step - loss: 0.2402 - accuracy: 0.9458 - val_loss: 0.8885 - val_accuracy: 0.8170\n",
            "Epoch 11/20\n",
            "16/16 [==============================] - 1s 52ms/step - loss: 0.2051 - accuracy: 0.9486 - val_loss: 0.9687 - val_accuracy: 0.7940\n",
            "Epoch 12/20\n",
            "16/16 [==============================] - 1s 53ms/step - loss: 0.1820 - accuracy: 0.9518 - val_loss: 0.9367 - val_accuracy: 0.8000\n",
            "Epoch 13/20\n",
            "16/16 [==============================] - 1s 52ms/step - loss: 0.1675 - accuracy: 0.9519 - val_loss: 0.9289 - val_accuracy: 0.8230\n",
            "Epoch 14/20\n",
            "16/16 [==============================] - 1s 51ms/step - loss: 0.1503 - accuracy: 0.9544 - val_loss: 1.0189 - val_accuracy: 0.8030\n",
            "Epoch 15/20\n",
            "16/16 [==============================] - 1s 66ms/step - loss: 0.1377 - accuracy: 0.9548 - val_loss: 0.9599 - val_accuracy: 0.8200\n",
            "Epoch 16/20\n",
            "16/16 [==============================] - 1s 52ms/step - loss: 0.1324 - accuracy: 0.9559 - val_loss: 1.0237 - val_accuracy: 0.8050\n",
            "Epoch 17/20\n",
            "16/16 [==============================] - 1s 52ms/step - loss: 0.1281 - accuracy: 0.9563 - val_loss: 0.9929 - val_accuracy: 0.8140\n",
            "Epoch 18/20\n",
            "16/16 [==============================] - 1s 51ms/step - loss: 0.1155 - accuracy: 0.9573 - val_loss: 1.0308 - val_accuracy: 0.8190\n",
            "Epoch 19/20\n",
            "16/16 [==============================] - 1s 52ms/step - loss: 0.1136 - accuracy: 0.9587 - val_loss: 1.0826 - val_accuracy: 0.7980\n",
            "Epoch 20/20\n",
            "16/16 [==============================] - 1s 52ms/step - loss: 0.1107 - accuracy: 0.9589 - val_loss: 1.1143 - val_accuracy: 0.7950\n"
          ]
        }
      ],
      "source": [
        "history = model.fit(partial_x_train,\n",
        "                    partial_y_train,\n",
        "                    epochs=20,\n",
        "                    batch_size=512,\n",
        "                    validation_data=(x_val, y_val))"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "IgfzW_8qRyea"
      },
      "source": [
        "**훈련과 검증 손실 그리기**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 32,
      "metadata": {
        "execution": {
          "iopub.execute_input": "2022-05-21T00:59:10.046902Z",
          "iopub.status.busy": "2022-05-21T00:59:10.045757Z",
          "iopub.status.idle": "2022-05-21T00:59:10.179653Z",
          "shell.execute_reply": "2022-05-21T00:59:10.178584Z"
        },
        "id": "SVm8Pe36Ryea",
        "outputId": "57221de9-5743-4e5a-d9ab-7e39caa4a1b4",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 295
        }
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ],
      "source": [
        "loss = history.history[\"loss\"]\n",
        "val_loss = history.history[\"val_loss\"]\n",
        "epochs = range(1, len(loss) + 1)\n",
        "plt.plot(epochs, loss, \"bo\", label=\"Training loss\")\n",
        "plt.plot(epochs, val_loss, \"b\", label=\"Validation loss\")\n",
        "plt.title(\"Training and validation loss\")\n",
        "plt.xlabel(\"Epochs\")\n",
        "plt.ylabel(\"Loss\")\n",
        "plt.legend()\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "rqBSs9lwRyea"
      },
      "source": [
        "**훈련과 검증 정확도 그리기**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 33,
      "metadata": {
        "execution": {
          "iopub.execute_input": "2022-05-21T00:59:10.183642Z",
          "iopub.status.busy": "2022-05-21T00:59:10.183348Z",
          "iopub.status.idle": "2022-05-21T00:59:10.318032Z",
          "shell.execute_reply": "2022-05-21T00:59:10.317232Z"
        },
        "id": "SNmU0tPzRyeb",
        "outputId": "4e800f19-a62c-4672-8ab5-09755d4aa823",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 295
        }
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ],
      "source": [
        "plt.clf()\n",
        "acc = history.history[\"accuracy\"]\n",
        "val_acc = history.history[\"val_accuracy\"]\n",
        "plt.plot(epochs, acc, \"bo\", label=\"Training accuracy\")\n",
        "plt.plot(epochs, val_acc, \"b\", label=\"Validation accuracy\")\n",
        "plt.title(\"Training and validation accuracy\")\n",
        "plt.xlabel(\"Epochs\")\n",
        "plt.ylabel(\"Accuracy\")\n",
        "plt.legend()\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "QIGN5jk2Ryeb"
      },
      "source": [
        "**모델을 처음부터 다시 훈련하기**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 34,
      "metadata": {
        "execution": {
          "iopub.execute_input": "2022-05-21T00:59:10.321952Z",
          "iopub.status.busy": "2022-05-21T00:59:10.321440Z",
          "iopub.status.idle": "2022-05-21T00:59:13.836469Z",
          "shell.execute_reply": "2022-05-21T00:59:13.835680Z"
        },
        "id": "DUaioa29Ryeb",
        "outputId": "d4191163-4a60-4d5b-8498-b3045ff8e983",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/9\n",
            "18/18 [==============================] - 2s 56ms/step - loss: 2.5854 - accuracy: 0.5176\n",
            "Epoch 2/9\n",
            "18/18 [==============================] - 1s 51ms/step - loss: 1.3371 - accuracy: 0.7187\n",
            "Epoch 3/9\n",
            "18/18 [==============================] - 1s 48ms/step - loss: 0.9913 - accuracy: 0.7925\n",
            "Epoch 4/9\n",
            "18/18 [==============================] - 1s 49ms/step - loss: 0.7797 - accuracy: 0.8361\n",
            "Epoch 5/9\n",
            "18/18 [==============================] - 1s 65ms/step - loss: 0.6174 - accuracy: 0.8696\n",
            "Epoch 6/9\n",
            "18/18 [==============================] - 1s 69ms/step - loss: 0.4938 - accuracy: 0.8979\n",
            "Epoch 7/9\n",
            "18/18 [==============================] - 1s 68ms/step - loss: 0.4011 - accuracy: 0.9156\n",
            "Epoch 8/9\n",
            "18/18 [==============================] - 1s 47ms/step - loss: 0.3256 - accuracy: 0.9286\n",
            "Epoch 9/9\n",
            "18/18 [==============================] - 1s 50ms/step - loss: 0.2741 - accuracy: 0.9385\n",
            "71/71 [==============================] - 1s 5ms/step - loss: 0.9717 - accuracy: 0.7818\n"
          ]
        }
      ],
      "source": [
        "model = keras.Sequential([\n",
        "  layers.Dense(64, activation=\"relu\"),\n",
        "  layers.Dense(64, activation=\"relu\"),\n",
        "  layers.Dense(46, activation=\"softmax\")\n",
        "])\n",
        "model.compile(optimizer=\"rmsprop\",\n",
        "              loss=\"categorical_crossentropy\",\n",
        "              metrics=[\"accuracy\"])\n",
        "model.fit(x_train,\n",
        "          y_train,\n",
        "          epochs=9,\n",
        "          batch_size=512)\n",
        "results = model.evaluate(x_test, y_test)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 35,
      "metadata": {
        "execution": {
          "iopub.execute_input": "2022-05-21T00:59:13.840276Z",
          "iopub.status.busy": "2022-05-21T00:59:13.839752Z",
          "iopub.status.idle": "2022-05-21T00:59:13.845156Z",
          "shell.execute_reply": "2022-05-21T00:59:13.844286Z"
        },
        "id": "8Uvx5UfiRyec",
        "outputId": "58f24210-fd8f-4d17-f6bc-6491bbc140a1",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[0.9716923236846924, 0.7818343639373779]"
            ]
          },
          "metadata": {},
          "execution_count": 35
        }
      ],
      "source": [
        "results"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 36,
      "metadata": {
        "execution": {
          "iopub.execute_input": "2022-05-21T00:59:13.848602Z",
          "iopub.status.busy": "2022-05-21T00:59:13.848317Z",
          "iopub.status.idle": "2022-05-21T00:59:13.854235Z",
          "shell.execute_reply": "2022-05-21T00:59:13.853462Z"
        },
        "id": "wiZnDW9HRyed",
        "outputId": "998a0c4e-48bf-4c89-fa46-3ccf582c9c8f",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "0.19011576135351738"
            ]
          },
          "metadata": {},
          "execution_count": 36
        }
      ],
      "source": [
        "import copy\n",
        "test_labels_copy = copy.copy(test_labels)\n",
        "np.random.shuffle(test_labels_copy)\n",
        "hits_array = np.array(test_labels) == np.array(test_labels_copy)\n",
        "hits_array.mean()"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Zt0QjrLiRyed"
      },
      "source": [
        "### 새로운 데이터로 예측하기"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 37,
      "metadata": {
        "execution": {
          "iopub.execute_input": "2022-05-21T00:59:13.858027Z",
          "iopub.status.busy": "2022-05-21T00:59:13.857648Z",
          "iopub.status.idle": "2022-05-21T00:59:14.132815Z",
          "shell.execute_reply": "2022-05-21T00:59:14.131779Z"
        },
        "id": "1rBJrITPRyee",
        "outputId": "120f1b2a-0d94-4804-f301-a1986d731f02",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "71/71 [==============================] - 0s 5ms/step\n"
          ]
        }
      ],
      "source": [
        "predictions = model.predict(x_test)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 38,
      "metadata": {
        "execution": {
          "iopub.execute_input": "2022-05-21T00:59:14.137393Z",
          "iopub.status.busy": "2022-05-21T00:59:14.136780Z",
          "iopub.status.idle": "2022-05-21T00:59:14.143161Z",
          "shell.execute_reply": "2022-05-21T00:59:14.142127Z"
        },
        "id": "oJm0-6OsRyee",
        "outputId": "d582093c-c99f-42a6-dc3c-76907cfac225",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(46,)"
            ]
          },
          "metadata": {},
          "execution_count": 38
        }
      ],
      "source": [
        "predictions[0].shape"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 39,
      "metadata": {
        "execution": {
          "iopub.execute_input": "2022-05-21T00:59:14.147504Z",
          "iopub.status.busy": "2022-05-21T00:59:14.146927Z",
          "iopub.status.idle": "2022-05-21T00:59:14.152400Z",
          "shell.execute_reply": "2022-05-21T00:59:14.151516Z"
        },
        "id": "iUAzDrwWRyef",
        "outputId": "4c734172-204a-4046-daa6-37dacbb5357f",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "0.9999999"
            ]
          },
          "metadata": {},
          "execution_count": 39
        }
      ],
      "source": [
        "np.sum(predictions[0])"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 40,
      "metadata": {
        "execution": {
          "iopub.execute_input": "2022-05-21T00:59:14.155905Z",
          "iopub.status.busy": "2022-05-21T00:59:14.155423Z",
          "iopub.status.idle": "2022-05-21T00:59:14.161331Z",
          "shell.execute_reply": "2022-05-21T00:59:14.160581Z"
        },
        "id": "fTJBpMyrRyef",
        "outputId": "0d6deb61-c0f8-46e5-b79a-9583bec77215",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "4"
            ]
          },
          "metadata": {},
          "execution_count": 40
        }
      ],
      "source": [
        "np.argmax(predictions[0])"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "7BVjZPk3Ryef"
      },
      "source": [
        "### 레이블과 손실을 다루는 다른 방법"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 41,
      "metadata": {
        "execution": {
          "iopub.execute_input": "2022-05-21T00:59:14.165413Z",
          "iopub.status.busy": "2022-05-21T00:59:14.164794Z",
          "iopub.status.idle": "2022-05-21T00:59:14.169329Z",
          "shell.execute_reply": "2022-05-21T00:59:14.168565Z"
        },
        "id": "eU3w1wAtRyef"
      },
      "outputs": [],
      "source": [
        "y_train = np.array(train_labels)\n",
        "y_test = np.array(test_labels)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 42,
      "metadata": {
        "execution": {
          "iopub.execute_input": "2022-05-21T00:59:14.173213Z",
          "iopub.status.busy": "2022-05-21T00:59:14.172609Z",
          "iopub.status.idle": "2022-05-21T00:59:14.181387Z",
          "shell.execute_reply": "2022-05-21T00:59:14.180511Z"
        },
        "id": "sUHGzOpeRyeg"
      },
      "outputs": [],
      "source": [
        "model.compile(optimizer=\"rmsprop\",\n",
        "              loss=\"sparse_categorical_crossentropy\",\n",
        "              metrics=[\"accuracy\"])"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "xoB6YZbLRyeh"
      },
      "source": [
        "### 충분히 큰 중간 층을 두어야 하는 이유"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "L1OwV1cfRyeh"
      },
      "source": [
        "**정보 병목이 있는 모델**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 43,
      "metadata": {
        "execution": {
          "iopub.execute_input": "2022-05-21T00:59:14.185530Z",
          "iopub.status.busy": "2022-05-21T00:59:14.185022Z",
          "iopub.status.idle": "2022-05-21T00:59:24.040585Z",
          "shell.execute_reply": "2022-05-21T00:59:24.039588Z"
        },
        "id": "E-_EV7p4Ryei",
        "outputId": "20182ad8-2edd-436d-b4b9-29490766b0b5",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/20\n",
            "63/63 [==============================] - 3s 27ms/step - loss: 2.9390 - accuracy: 0.2328 - val_loss: 2.3471 - val_accuracy: 0.2700\n",
            "Epoch 2/20\n",
            "63/63 [==============================] - 1s 23ms/step - loss: 1.8190 - accuracy: 0.5140 - val_loss: 1.5461 - val_accuracy: 0.6330\n",
            "Epoch 3/20\n",
            "63/63 [==============================] - 1s 20ms/step - loss: 1.2900 - accuracy: 0.6807 - val_loss: 1.3766 - val_accuracy: 0.6820\n",
            "Epoch 4/20\n",
            "63/63 [==============================] - 1s 19ms/step - loss: 1.1103 - accuracy: 0.7399 - val_loss: 1.2949 - val_accuracy: 0.7040\n",
            "Epoch 5/20\n",
            "63/63 [==============================] - 1s 19ms/step - loss: 0.9929 - accuracy: 0.7670 - val_loss: 1.2913 - val_accuracy: 0.7200\n",
            "Epoch 6/20\n",
            "63/63 [==============================] - 1s 24ms/step - loss: 0.9033 - accuracy: 0.7839 - val_loss: 1.2822 - val_accuracy: 0.7310\n",
            "Epoch 7/20\n",
            "63/63 [==============================] - 1s 22ms/step - loss: 0.8322 - accuracy: 0.7948 - val_loss: 1.2865 - val_accuracy: 0.7310\n",
            "Epoch 8/20\n",
            "63/63 [==============================] - 1s 19ms/step - loss: 0.7693 - accuracy: 0.8017 - val_loss: 1.3152 - val_accuracy: 0.7250\n",
            "Epoch 9/20\n",
            "63/63 [==============================] - 1s 23ms/step - loss: 0.7211 - accuracy: 0.8068 - val_loss: 1.3341 - val_accuracy: 0.7180\n",
            "Epoch 10/20\n",
            "63/63 [==============================] - 1s 24ms/step - loss: 0.6744 - accuracy: 0.8140 - val_loss: 1.3741 - val_accuracy: 0.7230\n",
            "Epoch 11/20\n",
            "63/63 [==============================] - 1s 21ms/step - loss: 0.6337 - accuracy: 0.8247 - val_loss: 1.4193 - val_accuracy: 0.7230\n",
            "Epoch 12/20\n",
            "63/63 [==============================] - 1s 20ms/step - loss: 0.5969 - accuracy: 0.8343 - val_loss: 1.4740 - val_accuracy: 0.7170\n",
            "Epoch 13/20\n",
            "63/63 [==============================] - 1s 19ms/step - loss: 0.5645 - accuracy: 0.8444 - val_loss: 1.5443 - val_accuracy: 0.7250\n",
            "Epoch 14/20\n",
            "63/63 [==============================] - 1s 19ms/step - loss: 0.5356 - accuracy: 0.8523 - val_loss: 1.5708 - val_accuracy: 0.7220\n",
            "Epoch 15/20\n",
            "63/63 [==============================] - 1s 19ms/step - loss: 0.5136 - accuracy: 0.8573 - val_loss: 1.6406 - val_accuracy: 0.7270\n",
            "Epoch 16/20\n",
            "63/63 [==============================] - 1s 19ms/step - loss: 0.4861 - accuracy: 0.8624 - val_loss: 1.6769 - val_accuracy: 0.7250\n",
            "Epoch 17/20\n",
            "63/63 [==============================] - 1s 23ms/step - loss: 0.4664 - accuracy: 0.8677 - val_loss: 1.7159 - val_accuracy: 0.7170\n",
            "Epoch 18/20\n",
            "63/63 [==============================] - 2s 25ms/step - loss: 0.4464 - accuracy: 0.8748 - val_loss: 1.7795 - val_accuracy: 0.7180\n",
            "Epoch 19/20\n",
            "63/63 [==============================] - 2s 24ms/step - loss: 0.4299 - accuracy: 0.8805 - val_loss: 1.8071 - val_accuracy: 0.7150\n",
            "Epoch 20/20\n",
            "63/63 [==============================] - 1s 23ms/step - loss: 0.4113 - accuracy: 0.8871 - val_loss: 1.8382 - val_accuracy: 0.7170\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<keras.callbacks.History at 0x7f855432dfd0>"
            ]
          },
          "metadata": {},
          "execution_count": 43
        }
      ],
      "source": [
        "model = keras.Sequential([\n",
        "    layers.Dense(64, activation=\"relu\"),\n",
        "    layers.Dense(4, activation=\"relu\"),\n",
        "    layers.Dense(46, activation=\"softmax\")\n",
        "])\n",
        "model.compile(optimizer=\"rmsprop\",\n",
        "              loss=\"categorical_crossentropy\",\n",
        "              metrics=[\"accuracy\"])\n",
        "model.fit(partial_x_train,\n",
        "          partial_y_train,\n",
        "          epochs=20,\n",
        "          batch_size=128,\n",
        "          validation_data=(x_val, y_val))"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "QRgFhIntRyei"
      },
      "source": [
        "### 추가 실험"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "BFtnILDjRyej"
      },
      "source": [
        "### 정리"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "f0v5NXFkRyej"
      },
      "source": [
        "## 주택 가격 예측: 회귀 문제"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "EiShhudRRyej"
      },
      "source": [
        "### 보스턴 주택 가격 데이터셋"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "PMOE4E-fRyek"
      },
      "source": [
        "**보스턴 주택 데이터셋 로드하기**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 44,
      "metadata": {
        "execution": {
          "iopub.execute_input": "2022-05-21T00:59:24.045461Z",
          "iopub.status.busy": "2022-05-21T00:59:24.044756Z",
          "iopub.status.idle": "2022-05-21T00:59:24.724575Z",
          "shell.execute_reply": "2022-05-21T00:59:24.723670Z"
        },
        "id": "9TVqKPRpRyek",
        "outputId": "44f6e809-d2f4-4f9d-dfd5-76d99873ab2e",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "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"
          ]
        }
      ],
      "source": [
        "from tensorflow.keras.datasets import boston_housing\n",
        "(train_data, train_targets), (test_data, test_targets) = boston_housing.load_data()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 45,
      "metadata": {
        "execution": {
          "iopub.execute_input": "2022-05-21T00:59:24.729145Z",
          "iopub.status.busy": "2022-05-21T00:59:24.728397Z",
          "iopub.status.idle": "2022-05-21T00:59:24.734773Z",
          "shell.execute_reply": "2022-05-21T00:59:24.733860Z"
        },
        "id": "p_5GRHojRyek",
        "outputId": "56da7d24-e65e-4ba0-f367-1f38e188ca81",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(404, 13)"
            ]
          },
          "metadata": {},
          "execution_count": 45
        }
      ],
      "source": [
        "train_data.shape"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 46,
      "metadata": {
        "execution": {
          "iopub.execute_input": "2022-05-21T00:59:24.738274Z",
          "iopub.status.busy": "2022-05-21T00:59:24.737701Z",
          "iopub.status.idle": "2022-05-21T00:59:24.743314Z",
          "shell.execute_reply": "2022-05-21T00:59:24.742477Z"
        },
        "id": "GtlvnaKPRyek",
        "outputId": "90e0a2de-a316-4977-fed1-a728a9b51afa",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(102, 13)"
            ]
          },
          "metadata": {},
          "execution_count": 46
        }
      ],
      "source": [
        "test_data.shape"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 47,
      "metadata": {
        "execution": {
          "iopub.execute_input": "2022-05-21T00:59:24.747014Z",
          "iopub.status.busy": "2022-05-21T00:59:24.746258Z",
          "iopub.status.idle": "2022-05-21T00:59:24.755724Z",
          "shell.execute_reply": "2022-05-21T00:59:24.754731Z"
        },
        "id": "Him5rrWrRyel",
        "outputId": "6546e3c2-650d-4ebd-8e0d-5671e77d6488",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([15.2, 42.3, 50. , 21.1, 17.7, 18.5, 11.3, 15.6, 15.6, 14.4, 12.1,\n",
              "       17.9, 23.1, 19.9, 15.7,  8.8, 50. , 22.5, 24.1, 27.5, 10.9, 30.8,\n",
              "       32.9, 24. , 18.5, 13.3, 22.9, 34.7, 16.6, 17.5, 22.3, 16.1, 14.9,\n",
              "       23.1, 34.9, 25. , 13.9, 13.1, 20.4, 20. , 15.2, 24.7, 22.2, 16.7,\n",
              "       12.7, 15.6, 18.4, 21. , 30.1, 15.1, 18.7,  9.6, 31.5, 24.8, 19.1,\n",
              "       22. , 14.5, 11. , 32. , 29.4, 20.3, 24.4, 14.6, 19.5, 14.1, 14.3,\n",
              "       15.6, 10.5,  6.3, 19.3, 19.3, 13.4, 36.4, 17.8, 13.5, 16.5,  8.3,\n",
              "       14.3, 16. , 13.4, 28.6, 43.5, 20.2, 22. , 23. , 20.7, 12.5, 48.5,\n",
              "       14.6, 13.4, 23.7, 50. , 21.7, 39.8, 38.7, 22.2, 34.9, 22.5, 31.1,\n",
              "       28.7, 46. , 41.7, 21. , 26.6, 15. , 24.4, 13.3, 21.2, 11.7, 21.7,\n",
              "       19.4, 50. , 22.8, 19.7, 24.7, 36.2, 14.2, 18.9, 18.3, 20.6, 24.6,\n",
              "       18.2,  8.7, 44. , 10.4, 13.2, 21.2, 37. , 30.7, 22.9, 20. , 19.3,\n",
              "       31.7, 32. , 23.1, 18.8, 10.9, 50. , 19.6,  5. , 14.4, 19.8, 13.8,\n",
              "       19.6, 23.9, 24.5, 25. , 19.9, 17.2, 24.6, 13.5, 26.6, 21.4, 11.9,\n",
              "       22.6, 19.6,  8.5, 23.7, 23.1, 22.4, 20.5, 23.6, 18.4, 35.2, 23.1,\n",
              "       27.9, 20.6, 23.7, 28. , 13.6, 27.1, 23.6, 20.6, 18.2, 21.7, 17.1,\n",
              "        8.4, 25.3, 13.8, 22.2, 18.4, 20.7, 31.6, 30.5, 20.3,  8.8, 19.2,\n",
              "       19.4, 23.1, 23. , 14.8, 48.8, 22.6, 33.4, 21.1, 13.6, 32.2, 13.1,\n",
              "       23.4, 18.9, 23.9, 11.8, 23.3, 22.8, 19.6, 16.7, 13.4, 22.2, 20.4,\n",
              "       21.8, 26.4, 14.9, 24.1, 23.8, 12.3, 29.1, 21. , 19.5, 23.3, 23.8,\n",
              "       17.8, 11.5, 21.7, 19.9, 25. , 33.4, 28.5, 21.4, 24.3, 27.5, 33.1,\n",
              "       16.2, 23.3, 48.3, 22.9, 22.8, 13.1, 12.7, 22.6, 15. , 15.3, 10.5,\n",
              "       24. , 18.5, 21.7, 19.5, 33.2, 23.2,  5. , 19.1, 12.7, 22.3, 10.2,\n",
              "       13.9, 16.3, 17. , 20.1, 29.9, 17.2, 37.3, 45.4, 17.8, 23.2, 29. ,\n",
              "       22. , 18. , 17.4, 34.6, 20.1, 25. , 15.6, 24.8, 28.2, 21.2, 21.4,\n",
              "       23.8, 31. , 26.2, 17.4, 37.9, 17.5, 20. ,  8.3, 23.9,  8.4, 13.8,\n",
              "        7.2, 11.7, 17.1, 21.6, 50. , 16.1, 20.4, 20.6, 21.4, 20.6, 36.5,\n",
              "        8.5, 24.8, 10.8, 21.9, 17.3, 18.9, 36.2, 14.9, 18.2, 33.3, 21.8,\n",
              "       19.7, 31.6, 24.8, 19.4, 22.8,  7.5, 44.8, 16.8, 18.7, 50. , 50. ,\n",
              "       19.5, 20.1, 50. , 17.2, 20.8, 19.3, 41.3, 20.4, 20.5, 13.8, 16.5,\n",
              "       23.9, 20.6, 31.5, 23.3, 16.8, 14. , 33.8, 36.1, 12.8, 18.3, 18.7,\n",
              "       19.1, 29. , 30.1, 50. , 50. , 22. , 11.9, 37.6, 50. , 22.7, 20.8,\n",
              "       23.5, 27.9, 50. , 19.3, 23.9, 22.6, 15.2, 21.7, 19.2, 43.8, 20.3,\n",
              "       33.2, 19.9, 22.5, 32.7, 22. , 17.1, 19. , 15. , 16.1, 25.1, 23.7,\n",
              "       28.7, 37.2, 22.6, 16.4, 25. , 29.8, 22.1, 17.4, 18.1, 30.3, 17.5,\n",
              "       24.7, 12.6, 26.5, 28.7, 13.3, 10.4, 24.4, 23. , 20. , 17.8,  7. ,\n",
              "       11.8, 24.4, 13.8, 19.4, 25.2, 19.4, 19.4, 29.1])"
            ]
          },
          "metadata": {},
          "execution_count": 47
        }
      ],
      "source": [
        "train_targets"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "X1yxfVb8Ryel"
      },
      "source": [
        "### 데이터 준비"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "8XtRErRzRyem"
      },
      "source": [
        "**데이터 정규화하기**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 48,
      "metadata": {
        "execution": {
          "iopub.execute_input": "2022-05-21T00:59:24.759692Z",
          "iopub.status.busy": "2022-05-21T00:59:24.759382Z",
          "iopub.status.idle": "2022-05-21T00:59:24.764808Z",
          "shell.execute_reply": "2022-05-21T00:59:24.763724Z"
        },
        "id": "XYrfTZ5MRyem"
      },
      "outputs": [],
      "source": [
        "mean = train_data.mean(axis=0)\n",
        "train_data -= mean\n",
        "std = train_data.std(axis=0)\n",
        "train_data /= std\n",
        "test_data -= mean\n",
        "test_data /= std"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "e3WGqjsFRyem"
      },
      "source": [
        "### 모델 구성"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "xWKDgk12Ryem"
      },
      "source": [
        "**모델 정의하기**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 49,
      "metadata": {
        "execution": {
          "iopub.execute_input": "2022-05-21T00:59:24.768846Z",
          "iopub.status.busy": "2022-05-21T00:59:24.768214Z",
          "iopub.status.idle": "2022-05-21T00:59:24.773358Z",
          "shell.execute_reply": "2022-05-21T00:59:24.772494Z"
        },
        "id": "MxJO2ZXlRyen"
      },
      "outputs": [],
      "source": [
        "def build_model():\n",
        "    model = keras.Sequential([\n",
        "        layers.Dense(64, activation=\"relu\"),\n",
        "        layers.Dense(64, activation=\"relu\"),\n",
        "        layers.Dense(1)\n",
        "    ])\n",
        "    model.compile(optimizer=\"rmsprop\", loss=\"mse\", metrics=[\"mae\"])\n",
        "    return model"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "P3KyOwjXRyen"
      },
      "source": [
        "### K-겹 검증을 사용한 훈련 검증"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Gvnz-7cIRyeo"
      },
      "source": [
        "**K-겹 검증하기**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 50,
      "metadata": {
        "execution": {
          "iopub.execute_input": "2022-05-21T00:59:24.777180Z",
          "iopub.status.busy": "2022-05-21T00:59:24.776509Z",
          "iopub.status.idle": "2022-05-21T00:59:35.954278Z",
          "shell.execute_reply": "2022-05-21T00:59:35.953206Z"
        },
        "id": "3YFddtBwRyeo",
        "outputId": "396b4bd4-77f7-4c94-edda-fb02443c26a5",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "#0번째 폴드 처리중\n",
            "#1번째 폴드 처리중\n",
            "#2번째 폴드 처리중\n",
            "#3번째 폴드 처리중\n"
          ]
        }
      ],
      "source": [
        "k = 4\n",
        "num_val_samples = len(train_data) // k\n",
        "num_epochs = 100\n",
        "all_scores = []\n",
        "for i in range(k):\n",
        "    print(f\"#{i}번째 폴드 처리중\")\n",
        "    val_data = train_data[i * num_val_samples: (i + 1) * num_val_samples]\n",
        "    val_targets = train_targets[i * num_val_samples: (i + 1) * num_val_samples]\n",
        "    partial_train_data = np.concatenate(\n",
        "        [train_data[:i * num_val_samples],\n",
        "         train_data[(i + 1) * num_val_samples:]],\n",
        "        axis=0)\n",
        "    partial_train_targets = np.concatenate(\n",
        "        [train_targets[:i * num_val_samples],\n",
        "         train_targets[(i + 1) * num_val_samples:]],\n",
        "        axis=0)\n",
        "    model = build_model()\n",
        "    model.fit(partial_train_data, partial_train_targets,\n",
        "              epochs=num_epochs, batch_size=16, verbose=0)\n",
        "    val_mse, val_mae = model.evaluate(val_data, val_targets, verbose=0)\n",
        "    all_scores.append(val_mae)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 51,
      "metadata": {
        "execution": {
          "iopub.execute_input": "2022-05-21T00:59:35.958698Z",
          "iopub.status.busy": "2022-05-21T00:59:35.958386Z",
          "iopub.status.idle": "2022-05-21T00:59:35.964375Z",
          "shell.execute_reply": "2022-05-21T00:59:35.963347Z"
        },
        "id": "vUv9Mo3yRyer",
        "outputId": "9a4f8bb7-225a-43b9-dba3-fe06e9dde5f7",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[1.9948174953460693, 2.536346435546875, 2.4903550148010254, 2.3979060649871826]"
            ]
          },
          "metadata": {},
          "execution_count": 51
        }
      ],
      "source": [
        "all_scores"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 52,
      "metadata": {
        "execution": {
          "iopub.execute_input": "2022-05-21T00:59:35.968402Z",
          "iopub.status.busy": "2022-05-21T00:59:35.967929Z",
          "iopub.status.idle": "2022-05-21T00:59:35.973477Z",
          "shell.execute_reply": "2022-05-21T00:59:35.972600Z"
        },
        "id": "e5AFRPuERyes",
        "outputId": "9efc08cb-c660-4650-d4a9-22f262c44516",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "2.354856252670288"
            ]
          },
          "metadata": {},
          "execution_count": 52
        }
      ],
      "source": [
        "np.mean(all_scores)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "AbzxHUVFRyes"
      },
      "source": [
        "**각 폴드의 검증 점수를 저장하기**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 53,
      "metadata": {
        "execution": {
          "iopub.execute_input": "2022-05-21T00:59:35.977004Z",
          "iopub.status.busy": "2022-05-21T00:59:35.976727Z",
          "iopub.status.idle": "2022-05-21T01:01:15.236431Z",
          "shell.execute_reply": "2022-05-21T01:01:15.235480Z"
        },
        "id": "tBqJ3nGjRyes",
        "outputId": "2ac55af6-0bf5-425f-f3ea-3341f636a5be",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "#0번째 폴드 처리중\n",
            "#1번째 폴드 처리중\n",
            "#2번째 폴드 처리중\n",
            "#3번째 폴드 처리중\n"
          ]
        }
      ],
      "source": [
        "num_epochs = 500\n",
        "all_mae_histories = []\n",
        "for i in range(k):\n",
        "    print(f\"#{i}번째 폴드 처리중\")\n",
        "    val_data = train_data[i * num_val_samples: (i + 1) * num_val_samples]\n",
        "    val_targets = train_targets[i * num_val_samples: (i + 1) * num_val_samples]\n",
        "    partial_train_data = np.concatenate(\n",
        "        [train_data[:i * num_val_samples],\n",
        "         train_data[(i + 1) * num_val_samples:]],\n",
        "        axis=0)\n",
        "    partial_train_targets = np.concatenate(\n",
        "        [train_targets[:i * num_val_samples],\n",
        "         train_targets[(i + 1) * num_val_samples:]],\n",
        "        axis=0)\n",
        "    model = build_model()\n",
        "    history = model.fit(partial_train_data, partial_train_targets,\n",
        "                        validation_data=(val_data, val_targets),\n",
        "                        epochs=num_epochs, batch_size=16, verbose=0)\n",
        "    mae_history = history.history[\"val_mae\"]\n",
        "    all_mae_histories.append(mae_history)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "n7hitqM_Ryet"
      },
      "source": [
        "**K-겹 검증 점수 평균을 기록하기**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 54,
      "metadata": {
        "execution": {
          "iopub.execute_input": "2022-05-21T01:01:15.240932Z",
          "iopub.status.busy": "2022-05-21T01:01:15.240614Z",
          "iopub.status.idle": "2022-05-21T01:01:15.249981Z",
          "shell.execute_reply": "2022-05-21T01:01:15.249150Z"
        },
        "id": "kt4SjG3NRyet"
      },
      "outputs": [],
      "source": [
        "average_mae_history = [\n",
        "    np.mean([x[i] for x in all_mae_histories]) for i in range(num_epochs)]"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "_TKKh0GDRyet"
      },
      "source": [
        "**검증 점수 그래프 그리기**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 55,
      "metadata": {
        "execution": {
          "iopub.execute_input": "2022-05-21T01:01:15.253642Z",
          "iopub.status.busy": "2022-05-21T01:01:15.253118Z",
          "iopub.status.idle": "2022-05-21T01:01:15.366352Z",
          "shell.execute_reply": "2022-05-21T01:01:15.365024Z"
        },
        "id": "pc47qQC4Ryet",
        "outputId": "c019a152-fc5d-4418-9302-266d6d956f34",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 279
        }
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ],
      "source": [
        "plt.plot(range(1, len(average_mae_history) + 1), average_mae_history)\n",
        "plt.xlabel(\"Epochs\")\n",
        "plt.ylabel(\"Validation MAE\")\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "WFRXT2qYRyet"
      },
      "source": [
        "**처음 10개의 데이터 포인트를 제외한 검증 점수 그래프 그리기**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 56,
      "metadata": {
        "execution": {
          "iopub.execute_input": "2022-05-21T01:01:15.371320Z",
          "iopub.status.busy": "2022-05-21T01:01:15.371023Z",
          "iopub.status.idle": "2022-05-21T01:01:15.481039Z",
          "shell.execute_reply": "2022-05-21T01:01:15.479606Z"
        },
        "id": "x6VFtaG4Ryeu",
        "outputId": "6d3701d1-eaf7-4970-d60c-acf6b63d65d0",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 279
        }
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ],
      "source": [
        "truncated_mae_history = average_mae_history[10:]\n",
        "plt.plot(range(1, len(truncated_mae_history) + 1), truncated_mae_history)\n",
        "plt.xlabel(\"Epochs\")\n",
        "plt.ylabel(\"Validation MAE\")\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "bxO2cIKpRyeu"
      },
      "source": [
        "**최종 모델 훈련하기**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 57,
      "metadata": {
        "execution": {
          "iopub.execute_input": "2022-05-21T01:01:15.488214Z",
          "iopub.status.busy": "2022-05-21T01:01:15.487884Z",
          "iopub.status.idle": "2022-05-21T01:01:20.038808Z",
          "shell.execute_reply": "2022-05-21T01:01:20.037674Z"
        },
        "id": "wSGcMeH3Ryev",
        "outputId": "c4c2b88f-1b65-4988-f8d2-776186e550ff",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "4/4 [==============================] - 0s 3ms/step - loss: 17.5936 - mae: 2.7210\n"
          ]
        }
      ],
      "source": [
        "model = build_model()\n",
        "model.fit(train_data, train_targets,\n",
        "          epochs=130, batch_size=16, verbose=0)\n",
        "test_mse_score, test_mae_score = model.evaluate(test_data, test_targets)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 58,
      "metadata": {
        "execution": {
          "iopub.execute_input": "2022-05-21T01:01:20.043289Z",
          "iopub.status.busy": "2022-05-21T01:01:20.042592Z",
          "iopub.status.idle": "2022-05-21T01:01:20.049022Z",
          "shell.execute_reply": "2022-05-21T01:01:20.048132Z"
        },
        "id": "_e_S3_P5Ryev",
        "outputId": "5b2c92ba-aee2-4078-a5b8-1a6964c6f21b",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "2.7209510803222656"
            ]
          },
          "metadata": {},
          "execution_count": 58
        }
      ],
      "source": [
        "test_mae_score"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "VesXPPJYRyev"
      },
      "source": [
        "### 새로운 데이터에 대해 예측하기"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 59,
      "metadata": {
        "execution": {
          "iopub.execute_input": "2022-05-21T01:01:20.053135Z",
          "iopub.status.busy": "2022-05-21T01:01:20.052814Z",
          "iopub.status.idle": "2022-05-21T01:01:20.154058Z",
          "shell.execute_reply": "2022-05-21T01:01:20.153234Z"
        },
        "id": "rWLRtftSRyew",
        "outputId": "6fb22d99-d0ad-44c0-dce6-772f846541c2",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "4/4 [==============================] - 0s 4ms/step\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([6.934491], dtype=float32)"
            ]
          },
          "metadata": {},
          "execution_count": 59
        }
      ],
      "source": [
        "predictions = model.predict(test_data)\n",
        "predictions[0]"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "cyvfquV2Ryez"
      },
      "source": [
        "### 정리"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "hN6PNjwSRyez"
      },
      "source": [
        "## 요약"
      ]
    }
  ],
  "metadata": {
    "colab": {
      "collapsed_sections": [],
      "name": "chapter04_getting-started-with-neural-networks.i",
      "provenance": []
    },
    "kernelspec": {
      "display_name": "default:Python",
      "language": "python",
      "name": "conda-env-default-py"
    },
    "language_info": {
      "codemirror_mode": {
        "name": "ipython",
        "version": 3
      },
      "file_extension": ".py",
      "mimetype": "text/x-python",
      "name": "python",
      "nbconvert_exporter": "python",
      "pygments_lexer": "ipython3",
      "version": "3.9.10"
    }
  },
  "nbformat": 4,
  "nbformat_minor": 0
}