neural_network_and_classifying_movie_reviews.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "neural network and classifying movie reviews.ipynb",
      "provenance": [],
      "collapsed_sections": [],
      "authorship_tag": "ABX9TyNfpnOVeCrZZpoS1tHiXX6v",
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/JSY0717/Deep-Learning-and-Application/blob/main/neural_network_and_classifying_movie_reviews.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "e_2FiiLK8JFm"
      },
      "source": [
        "# 신경망을 위한 데이터 표현"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "fIKgMS2O6UU2"
      },
      "source": [
        "import tensorflow\r\n",
        "import keras"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 36
        },
        "id": "UQg4mguM6dIY",
        "outputId": "889bf3f0-501f-428c-b25c-bbd218839e3a"
      },
      "source": [
        "keras.__version__"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "application/vnd.google.colaboratory.intrinsic+json": {
              "type": "string"
            },
            "text/plain": [
              "'2.4.3'"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 106
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "S7Un2poK8P5j"
      },
      "source": [
        "### • 1) 손글씨 데이터 분류 (MNIST dataset)"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "7A1WJ8EE6mVi"
      },
      "source": [
        "from keras.datasets import mnist\r\n",
        "\r\n",
        "(train_images, train_labels), (test_images, test_labels) = mnist.load_data()"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "yAQGehmV6oAC",
        "outputId": "1ff1fd96-50f2-4e31-8558-374306219340"
      },
      "source": [
        "print(train_images.shape)\r\n",
        "print(train_images.ndim)\r\n",
        "print(train_images.dtype)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "(60000, 28, 28)\n",
            "3\n",
            "uint8\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "cqfvKamr6t1W",
        "outputId": "9b20cf3b-89dd-47bc-cc32-dae6eec2e549"
      },
      "source": [
        "len(train_labels) "
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "60000"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 109
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ewehWQCm0Qp1",
        "outputId": "03511d43-62f7-4a81-9667-5a386178a721"
      },
      "source": [
        "train_labels"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([5, 0, 4, ..., 5, 6, 8], dtype=uint8)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 110
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "EuTX_OWS0VpN",
        "outputId": "fc74c93e-888d-4da3-db0b-b7dd84465950"
      },
      "source": [
        "test_images.shape"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(10000, 28, 28)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 111
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 265
        },
        "id": "6drF9Vsr66i4",
        "outputId": "202c5cf7-d2ef-4fa6-b05b-d86d89e23a4c"
      },
      "source": [
        "# 3D 텐서에서 matplotlib을 사용\r\n",
        "import matplotlib.pyplot as plt\r\n",
        "\r\n",
        "digit = train_images[4]\r\n",
        "plt.imshow(digit, cmap=plt.cm.binary)\r\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 761
        },
        "id": "zKz6oHn3uizz",
        "outputId": "f7e817ae-03fa-4de0-8501-999522e77995"
      },
      "source": [
        "numbers=[0,1000,10000]\r\n",
        "\r\n",
        "for i in numbers:  # while i<3\r\n",
        "  digit = train_images[i]\r\n",
        "  i=i+1\r\n",
        "  plt.imshow(digit, cmap=plt.cm.binary)\r\n",
        "  plt.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "XYfttFNNH7Jr"
      },
      "source": [
        "*+) numpy로 tensor 조작*\r\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "XGbNNmBDHJTk",
        "outputId": "c2ef67d2-9ebf-4179-f812-3c1ce87a2607"
      },
      "source": [
        "# slicing\r\n",
        "my_sl = train_images[10:100]\r\n",
        "#                   [10:100, 0:28, 0:28] 과 동일\r\n",
        "my_sl.shape"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(90, 28, 28)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 8
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "MZ6buAqJIkW-"
      },
      "source": [
        "*+) 배치 데이터 : tensor의 첫번째 축을 작게 나눈 단위*\r\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "j_MOIa-1IFkj"
      },
      "source": [
        "n=0\r\n",
        "batch_n = train_images[128*n : 128*(n+1)]"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "he7NUm8otrY3"
      },
      "source": [
        "### 1-1) 신경망 생성"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "6P8kFgdZvjcf"
      },
      "source": [
        "신경망은 데이터 처리 필터 역할을 하는 '층' 으로 구성되어 있다. 데이터가 층을 통과하면서 점진적으로 정제된다.  \r\n",
        "딥러닝 모델은 데이터 정제 필터가 연속된 여과기와 같다.  \r\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "WisuZs-stk2S"
      },
      "source": [
        "from keras import models,layers\r\n",
        "\r\n",
        "network = models.Sequential()\r\n",
        "network.add(layers.Dense(512, activation='relu', input_shape=(28 * 28,)))\r\n",
        "network.add(layers.Dense(10, activation='softmax'))"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "BPRqiwXHv_mF"
      },
      "source": [
        ".add() 를 통해 Dense 신경망 층 2개를 연속되게 했다. 각 점수는 현재 숫자 이미지가 10개의 숫자 클래스 중 하나에 속할 확률이다."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "d4cDw6QSJTjK",
        "outputId": "3fc03a9c-8818-4563-ff63-2fa7b43a9265"
      },
      "source": [
        "network.summary()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Model: \"sequential_5\"\n",
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "dense_13 (Dense)             (None, 512)               401920    \n",
            "_________________________________________________________________\n",
            "dense_14 (Dense)             (None, 10)                5130      \n",
            "=================================================================\n",
            "Total params: 407,050\n",
            "Trainable params: 407,050\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "gdnPKmc-wnot"
      },
      "source": [
        "신경망 컴파일 단계에서 필요한 세 파라미터는 다음과 같다.  \r\n",
        "\r\n",
        "1. 손실함수 : train data에서 신경망의 성능 측정  \r\n",
        "2. 옵티마이저 : 입력된 데이터와 손실 함수를 기반으로 네트워크를 업데이트하는 메커니즘  \r\n",
        "3. 훈련과 테스트 과정을 모니터링할 지표 : 여기에서는 정확도"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "_Kw_at9tvX6q"
      },
      "source": [
        "network.compile(optimizer='rmsprop',\r\n",
        "                loss='categorical_crossentropy',\r\n",
        "                metrics=['accuracy'])"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "pKOVe4z1xVJn"
      },
      "source": [
        "# 전처리\r\n",
        "# 1) 데이터를 network에 맞는 크기로 바꾸고, 모든 값을 0과 1 사이로 scale 조정\r\n",
        "# 0~255 -> 0~1\r\n",
        "\r\n",
        "train_images = train_images.reshape((60000, 28 * 28))\r\n",
        "train_images = train_images.astype('float32') / 255\r\n",
        "\r\n",
        "test_images = test_images.reshape((10000, 28 * 28))\r\n",
        "test_images = test_images.astype('float32') / 255\r\n",
        "\r\n",
        "# 2) 레이블을 범주형으로 인코딩\r\n",
        "from keras.utils import to_categorical\r\n",
        "\r\n",
        "train_labels = to_categorical(train_labels)\r\n",
        "test_labels = to_categorical(test_labels)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "oD2sHYO2x5uG",
        "outputId": "e4154938-a6a6-4cfd-c84a-9d7233cc833b"
      },
      "source": [
        "# .fit() 메서드를 통해 train data에 모델을 학습\r\n",
        "network.fit(train_images, train_labels, epochs=5, batch_size=128)\r\n",
        "\r\n",
        "#약 98.9%의 높은 정확도"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Epoch 1/5\n",
            "469/469 [==============================] - 3s 7ms/step - loss: 0.4187 - accuracy: 0.8802\n",
            "Epoch 2/5\n",
            "469/469 [==============================] - 3s 7ms/step - loss: 0.1053 - accuracy: 0.9685\n",
            "Epoch 3/5\n",
            "469/469 [==============================] - 3s 7ms/step - loss: 0.0697 - accuracy: 0.9792\n",
            "Epoch 4/5\n",
            "469/469 [==============================] - 3s 7ms/step - loss: 0.0480 - accuracy: 0.9855\n",
            "Epoch 5/5\n",
            "469/469 [==============================] - 3s 7ms/step - loss: 0.0371 - accuracy: 0.9892\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<tensorflow.python.keras.callbacks.History at 0x7fbf12c65240>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 116
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "zn21llLrz3HQ",
        "outputId": "f44f96b0-c52e-45ee-94ac-17e5ec695778"
      },
      "source": [
        "# test data에 network 적용\r\n",
        "test_loss, test_acc = network.evaluate(test_images, test_labels)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "313/313 [==============================] - 1s 2ms/step - loss: 0.0678 - accuracy: 0.9802\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "41slZKjU0Bs2",
        "outputId": "79e28dd7-8389-4070-96c7-782c87de7217"
      },
      "source": [
        "print('test_acc:', test_acc)\r\n",
        "# 98.0%로, train data에서보다는 낮음. 차이가 발생하는 이유는 과대적합 때문."
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "test_acc: 0.9801999926567078\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "9OH09to2TF5X"
      },
      "source": [
        "### •  2) 영화 리뷰 분류 (IMDB dataset)"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "o-s3pKXkTObr",
        "outputId": "1cfe4ff5-3752-46fb-c023-9f24fa784646"
      },
      "source": [
        "# imdb dataset 로드\r\n",
        "from keras.datasets import imdb\r\n",
        "\r\n",
        "# traindata, testdata 는 리뷰의 목록이다. 각 리뷰는 단어가 아니라 단어 인덱스의 리스트이다.\r\n",
        "# label 은 리뷰가 긍정적인지, 부정적인지를 나타내는 0 or 1\r\n",
        "(traindata,trainlabel),(testdata,testlabel) = imdb.load_data(num_words=10000)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "<string>:6: VisibleDeprecationWarning: Creating an ndarray from ragged nested sequences (which is a list-or-tuple of lists-or-tuples-or ndarrays with different lengths or shapes) is deprecated. If you meant to do this, you must specify 'dtype=object' when creating the ndarray\n",
            "/usr/local/lib/python3.6/dist-packages/tensorflow/python/keras/datasets/imdb.py:159: VisibleDeprecationWarning: Creating an ndarray from ragged nested sequences (which is a list-or-tuple of lists-or-tuples-or ndarrays with different lengths or shapes) is deprecated. If you meant to do this, you must specify 'dtype=object' when creating the ndarray\n",
            "  x_train, y_train = np.array(xs[:idx]), np.array(labels[:idx])\n",
            "/usr/local/lib/python3.6/dist-packages/tensorflow/python/keras/datasets/imdb.py:160: VisibleDeprecationWarning: Creating an ndarray from ragged nested sequences (which is a list-or-tuple of lists-or-tuples-or ndarrays with different lengths or shapes) is deprecated. If you meant to do this, you must specify 'dtype=object' when creating the ndarray\n",
            "  x_test, y_test = np.array(xs[idx:]), np.array(labels[idx:])\n"
          ],
          "name": "stderr"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "HlyuI3EanfYo"
      },
      "source": [
        "traindata[0]"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "xWKunZQlTu0u",
        "outputId": "552458b7-1235-4e84-ca6a-0c6a0cf00b33"
      },
      "source": [
        "# 리뷰데이터 하나를 원래 영어단어로 바꾸는 예제 (단어 인덱스->단어들)\r\n",
        "\r\n",
        "# 단어와 정수 인덱스를 매핑한 딕셔너리 호출\r\n",
        "word_index = imdb.get_word_index()\r\n",
        "# 정수 인덱스와 단어를 매핑하도록 뒤집는다\r\n",
        "reverse_word_index = dict([(value,key) for (key,value) in word_index.items()])\r\n",
        "# 영어단어로 변환\r\n",
        "decoded_review= ' '.join([reverse_word_index.get(i - 3,'?') for i in traindata[0]])"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Downloading data from https://storage.googleapis.com/tensorflow/tf-keras-datasets/imdb_word_index.json\n",
            "1646592/1641221 [==============================] - 0s 0us/step\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 148
        },
        "id": "U0wEfEjoUjIN",
        "outputId": "a268aa6f-f34a-4fce-f879-b57cfa09f0b4"
      },
      "source": [
        "decoded_review"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "application/vnd.google.colaboratory.intrinsic+json": {
              "type": "string"
            },
            "text/plain": [
              "\"? this film was just brilliant casting location scenery story direction everyone's really suited the part they played and you could just imagine being there robert ? is an amazing actor and now the same being director ? father came from the same scottish island as myself so i loved the fact there was a real connection with this film the witty remarks throughout the film were great it was just brilliant so much that i bought the film as soon as it was released for ? and would recommend it to everyone to watch and the fly fishing was amazing really cried at the end it was so sad and you know what they say if you cry at a film it must have been good and this definitely was also ? to the two little boy's that played the ? of norman and paul they were just brilliant children are often left out of the ? list i think because the stars that play them all grown up are such a big profile for the whole film but these children are amazing and should be praised for what they have done don't you think the whole story was so lovely because it was true and was someone's life after all that was shared with us all\""
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 12
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "XzsWmd3rcnQV"
      },
      "source": [
        "### 2-1) 데이터 준비"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "F5Uc1uduXKQ9"
      },
      "source": [
        "신경망에 숫자 리스트를 입력할 수 없기 때문에, 리스트를 텐서로 변경해야 한다.\r\n",
        "\r\n",
        "방법 1) 같은 길이가 되도록 리스트에 패딩을 추가하고, (saples, sequence_length) 크기의 정수 텐서로 변환  \r\n",
        "방법 2) one-hot enconding"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "QZqZh21wV1Pv"
      },
      "source": [
        "# 방법 2) 정수 시퀀스를 이진행렬로 인코딩\r\n",
        "import numpy as np\r\n",
        "\r\n",
        "def vectorize_sequences(sequences, dimension=10000):\r\n",
        "  results = np.zeros((len(sequences),dimension)) # 영행렬 생성\r\n",
        "  for i,sequences in enumerate (sequences):\r\n",
        "    results[i, sequences] = 1. # 특정 인덱스의 위치를 1로 update\r\n",
        "    return results\r\n",
        "\r\n",
        "xtrain = vectorize_sequences(traindata)\r\n",
        "xtest = vectorize_sequences(testdata)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "4mneid1BZ6xg",
        "outputId": "21b236b1-d6ea-43c5-80eb-b044c3b5ad56"
      },
      "source": [
        "xtrain[0]"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([0., 1., 1., ..., 0., 0., 0.])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 14
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Xsj-NCeuaA41"
      },
      "source": [
        "# 레이블을 벡터로 변환 (이미 1과 0으로 구성되어 있으므로 one-hot enconding 불필요)\r\n",
        "\r\n",
        "ytrain = np.asarray(trainlabel).astype('float32')\r\n",
        "ytest = np.asarray(testlabel).astype('float32')"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "BzdxaiI8cU4a"
      },
      "source": [
        "### 2-2) 신경망 모델 만들기"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "JCmQ4iLBb4eT"
      },
      "source": [
        "from keras import models,layers\r\n",
        "\r\n",
        "mymodel = models.Sequential()\r\n",
        "mymodel.add(layers.Dense(16, activation='relu',input_shape=(10000,)))  # 16개의 은닉층\r\n",
        "mymodel.add(layers.Dense(16, activation='relu'))\r\n",
        "mymodel.add(layers.Dense(1, activation='sigmoid'))"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "2QkVQJgRfkNK",
        "outputId": "519a9ba1-85b1-48f6-be71-fe1dc059ef6a"
      },
      "source": [
        "mymodel.summary()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Model: \"sequential\"\n",
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "dense (Dense)                (None, 16)                160016    \n",
            "_________________________________________________________________\n",
            "dense_1 (Dense)              (None, 16)                272       \n",
            "_________________________________________________________________\n",
            "dense_2 (Dense)              (None, 1)                 17        \n",
            "=================================================================\n",
            "Total params: 160,305\n",
            "Trainable params: 160,305\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "upHeoNp7fzDG"
      },
      "source": [
        "# 모델 컴파일\r\n",
        "# 방법 1)\r\n",
        "\r\n",
        "mymodel.compile(optimizer='rmsprop',\r\n",
        "                loss='binary_crossentropy',\r\n",
        "                metrics=['accuracy'])"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "n1_gzwDhk37P"
      },
      "source": [
        "# 모델 컴파일\r\n",
        "# 방법 2) 옵티마이저의 매개변수 설정하기\r\n",
        "#         loss와 metrics를 함수 객체로 지정하기\r\n",
        "\r\n",
        "from keras import optimizers\r\n",
        "\r\n",
        "mymodel.compile(optimizer=optimizers.RMSprop(lr=0.001),\r\n",
        "                loss='binary_crossentropy',\r\n",
        "                metrics=['accuracy'])\r\n",
        "\r\n",
        "from keras import losses,metrics\r\n",
        "\r\n",
        "mymodel.compile(optimizer=optimizers.RMSprop(lr=0.001),\r\n",
        "                loss=losses.binary_crossentropy,\r\n",
        "                metrics=[metrics.binary_accuracy])"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "J0913PEpnjyl"
      },
      "source": [
        "### 2-3) 훈련 검증"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "h38FwfZjlPVL"
      },
      "source": [
        "# 검증 데이터(처음 본 데이터) 준비\r\n",
        "\r\n",
        "xvalid = xtrain[:10000]\r\n",
        "partial_xtrain = xtrain[10000:]\r\n",
        "\r\n",
        "yvalid = ytrain[:10000]\r\n",
        "partial_ytrain = ytrain[10000:]"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "TWnxvDuIln51",
        "outputId": "5395ec68-afaa-4b39-99c8-05f840f14095"
      },
      "source": [
        "partial_xtrain"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([[0., 0., 0., ..., 0., 0., 0.],\n",
              "       [0., 0., 0., ..., 0., 0., 0.],\n",
              "       [0., 0., 0., ..., 0., 0., 0.],\n",
              "       ...,\n",
              "       [0., 0., 0., ..., 0., 0., 0.],\n",
              "       [0., 0., 0., ..., 0., 0., 0.],\n",
              "       [0., 0., 0., ..., 0., 0., 0.]])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 50
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "h2dBKLZLm8QK",
        "outputId": "9a51001c-fa5f-43c6-aaee-6fe875fe5138"
      },
      "source": [
        "# 모델을 512개의 sample 단위로 미니 batch를 만들어,20번의 에포크동안 훈련\r\n",
        "# (xtrain과 ytrain tensor에 있는 모든 샘플에 대해 20번 반복)\r\n",
        "# ,fit() : 훈련하는 동안 발생한 모든 정보를 담고 있는 딕셔너리인 history 객체를 반환\r\n",
        "\r\n",
        "history = mymodel.fit(partial_xtrain,\r\n",
        "                    partial_ytrain,\r\n",
        "                    epochs=20,\r\n",
        "                    batch_size=512,\r\n",
        "                    validation_data=(xvalid,yvalid))"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Epoch 1/20\n",
            "30/30 [==============================] - 2s 55ms/step - loss: 0.6932 - accuracy: 0.4991 - val_loss: 0.6932 - val_accuracy: 0.4948\n",
            "Epoch 2/20\n",
            "30/30 [==============================] - 1s 30ms/step - loss: 0.6932 - accuracy: 0.4891 - val_loss: 0.6932 - val_accuracy: 0.4947\n",
            "Epoch 3/20\n",
            "30/30 [==============================] - 1s 32ms/step - loss: 0.6932 - accuracy: 0.5007 - val_loss: 0.6932 - val_accuracy: 0.4947\n",
            "Epoch 4/20\n",
            "30/30 [==============================] - 1s 32ms/step - loss: 0.6931 - accuracy: 0.5032 - val_loss: 0.6932 - val_accuracy: 0.4947\n",
            "Epoch 5/20\n",
            "30/30 [==============================] - 1s 30ms/step - loss: 0.6931 - accuracy: 0.5081 - val_loss: 0.6932 - val_accuracy: 0.4947\n",
            "Epoch 6/20\n",
            "30/30 [==============================] - 1s 31ms/step - loss: 0.6932 - accuracy: 0.4990 - val_loss: 0.6932 - val_accuracy: 0.4947\n",
            "Epoch 7/20\n",
            "30/30 [==============================] - 1s 30ms/step - loss: 0.6931 - accuracy: 0.5061 - val_loss: 0.6932 - val_accuracy: 0.4947\n",
            "Epoch 8/20\n",
            "30/30 [==============================] - 1s 30ms/step - loss: 0.6932 - accuracy: 0.4995 - val_loss: 0.6933 - val_accuracy: 0.4947\n",
            "Epoch 9/20\n",
            "30/30 [==============================] - 1s 30ms/step - loss: 0.6931 - accuracy: 0.5060 - val_loss: 0.6933 - val_accuracy: 0.4947\n",
            "Epoch 10/20\n",
            "30/30 [==============================] - 1s 31ms/step - loss: 0.6930 - accuracy: 0.5121 - val_loss: 0.6932 - val_accuracy: 0.4947\n",
            "Epoch 11/20\n",
            "30/30 [==============================] - 1s 31ms/step - loss: 0.6931 - accuracy: 0.5084 - val_loss: 0.6932 - val_accuracy: 0.4947\n",
            "Epoch 12/20\n",
            "30/30 [==============================] - 1s 30ms/step - loss: 0.6931 - accuracy: 0.5076 - val_loss: 0.6932 - val_accuracy: 0.4947\n",
            "Epoch 13/20\n",
            "30/30 [==============================] - 1s 31ms/step - loss: 0.6932 - accuracy: 0.4999 - val_loss: 0.6932 - val_accuracy: 0.4947\n",
            "Epoch 14/20\n",
            "30/30 [==============================] - 1s 31ms/step - loss: 0.6931 - accuracy: 0.5023 - val_loss: 0.6932 - val_accuracy: 0.4947\n",
            "Epoch 15/20\n",
            "30/30 [==============================] - 1s 30ms/step - loss: 0.6933 - accuracy: 0.4926 - val_loss: 0.6933 - val_accuracy: 0.4947\n",
            "Epoch 16/20\n",
            "30/30 [==============================] - 1s 31ms/step - loss: 0.6932 - accuracy: 0.4988 - val_loss: 0.6933 - val_accuracy: 0.4947\n",
            "Epoch 17/20\n",
            "30/30 [==============================] - 1s 30ms/step - loss: 0.6932 - accuracy: 0.4957 - val_loss: 0.6932 - val_accuracy: 0.4947\n",
            "Epoch 18/20\n",
            "30/30 [==============================] - 1s 30ms/step - loss: 0.6930 - accuracy: 0.5089 - val_loss: 0.6932 - val_accuracy: 0.4947\n",
            "Epoch 19/20\n",
            "30/30 [==============================] - 1s 30ms/step - loss: 0.6931 - accuracy: 0.5038 - val_loss: 0.6932 - val_accuracy: 0.4947\n",
            "Epoch 20/20\n",
            "30/30 [==============================] - 1s 29ms/step - loss: 0.6931 - accuracy: 0.5032 - val_loss: 0.6932 - val_accuracy: 0.4947\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "mN7ADuXlhGSF",
        "outputId": "48f3ae0a-7e56-4506-c872-e185578a91ac"
      },
      "source": [
        "history_dict = history.history\r\n",
        "history_dict.keys()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "dict_keys(['loss', 'accuracy', 'val_loss', 'val_accuracy'])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 65
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "aqePkuthgDr1"
      },
      "source": [
        "### 2-4) matplotlib을 활용해 훈련과 검증 데이터에 대한 loss와 accuracy 시각화"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "hZqmI1Aqhq0E"
      },
      "source": [
        "import matplotlib.pyplot as plt"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 295
        },
        "id": "2H118Z7OdEVJ",
        "outputId": "1f2b48b6-20bd-4232-ad23-5e0dabe785a1"
      },
      "source": [
        "#loss\r\n",
        "\r\n",
        "loss = history_dict['loss']\r\n",
        "val_loss = history_dict['val_loss']\r\n",
        "\r\n",
        "epochs = range(1,len(acc)+1)\r\n",
        "\r\n",
        "plt.plot(epochs, loss, 'bo', label='Training loss')\r\n",
        "plt.plot(epochs, val_loss, 'b', label='Validation loss')\r\n",
        "plt.title('Training and Validation Loss')\r\n",
        "plt.xlabel('Epochs')\r\n",
        "plt.ylabel('Loss')\r\n",
        "plt.legend()\r\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAZsAAAEWCAYAAACwtjr+AAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nO3dd5xU1fnH8c+XJi5FqkZBASNoUKkLFixYg2IgYiVYEMWowViiEX8kQqxR0SiKSVDABoKVYAUlFmyJKwIKoiIiYiwICiggLDy/P85dHJbdZcvcubO7z/v1mtfOnLnluTOz88w599xzZGY455xzcaqRdADOOeeqPk82zjnnYufJxjnnXOw82TjnnIudJxvnnHOx82TjnHMudp5sXKUi6VlJZ6Z72SRJWizpyBi2+5Kkc6L7AyRNL82y5djPbpK+l1SzvLG6qs+TjYtd9EVUcNskaW3K4wFl2ZaZHWNm96V72WwkaaikV4oobyZpvaR9SrstM5tgZkenKa4tkqOZLTGz+ma2MR3bL7Qvk7RHurfrMs+TjYtd9EVU38zqA0uAX6WUTShYTlKt5KLMSg8CB0pqU6j8VOBdM3svgZicKxdPNi4xknpKWirpCklfAuMlNZb0lKRlkr6N7rdMWSe1aWigpFcljYyW/UTSMeVcto2kVyStlvSCpNGSHiwm7tLEeI2k16LtTZfULOX50yV9Kmm5pGHFvT5mthT4N3B6oafOAO7fVhyFYh4o6dWUx0dJWiBppaQ7AaU893NJ/47i+0bSBEmNouceAHYDnoxqpn+U1DqqgdSKltlF0lRJKyQtlDQ4ZdsjJD0s6f7otZknKbe416A4knaItrEsei3/JKlG9Nwekl6Oju0bSZOjckn6m6SvJa2S9G5ZaoeuYjzZuKT9DGgCtALOJXwmx0ePdwPWAneWsP5+wAdAM+AmYKwklWPZicB/gabACLb+gk9Vmhh/A5wF7AjUAS4DkNQe+Hu0/V2i/RWZICL3pcYiaU+gUxRvWV+rgm00Ax4H/kR4LT4GeqQuAtwQxfcLYFfCa4KZnc6WtdObitjFJGBptP6JwPWSDk95vk+0TCNgamliLsIdwA7A7sChhAR8VvTcNcB0oDHhtb0jKj8aOARoF617MrC8HPt25WFmfvNbxm7AYuDI6H5PYD1Qt4TlOwHfpjx+CTgnuj8QWJjyXA5gwM/KsizhizofyEl5/kHgwVIeU1Ex/inl8QXAc9H9q4BJKc/Vi16DI4vZdg6wCjgwenwd8K9yvlavRvfPAN5MWU6E5HBOMdv9NfBOUe9h9Lh19FrWIiSmjUCDlOdvAO6N7o8AXkh5rj2wtoTX1oA9CpXVjF6z9illvwVeiu7fD4wBWhZa73DgQ2B/oEbS/wvV7eY1G5e0ZWa2ruCBpBxJ/4yaRlYBrwCNVHxPpy8L7pjZmuhu/TIuuwuwIqUM4LPiAi5ljF+m3F+TEtMuqds2sx8o4dd1FNMjwBlRLWwA4cu0PK9VgcIxWOpjSTtJmiTp82i7DxJqQKVR8FquTin7FGiR8rjwa1NXZTtf1wyoHW23qH38kZBA/xs10w0CMLN/E2pRo4GvJY2R1LAM+3UV4MnGJa3wsON/APYE9jOzhoRmD0g5pxCDL4AmknJSynYtYfmKxPhF6rajfTbdxjr3EZp8jgIaAE9WMI7CMYgtj/d6wvuyb7Td0wpts6Sh4v9HeC0bpJTtBny+jZjK4htgA6H5cKt9mNmXZjbYzHYh1HjuUtSjzcxGmVlXQo2qHXB5GuNyJfBk47JNA8K5h+8kNQGGx71DM/sUyANGSKoj6QDgVzHF+ChwnKSDJNUBrmbb/4czge8ITUOTzGx9BeN4GthbUr+oRvF7QnNigQbA98BKSS3Y+gv5K8K5kq2Y2WfA68ANkupK6gCcTagdlVedaFt1JdWNyh4GrpPUQFIr4NKCfUg6KaWjxLeE5LhJUjdJ+0mqDfwArAM2VSAuVwaebFy2uQ3YnvDr9U3guQztdwBwAKFJ61pgMvBjMcuWO0Yzmwf8jnCC/wvCl+HSbaxjhKazVtHfCsVhZt8AJwF/JRxvW+C1lEX+AnQBVhIS0+OFNnED8CdJ30m6rIhd9Cecx/kf8AQw3MxeKE1sxZhHSKoFt7OACwkJYxHwKuH1HBct3w34j6TvCR0QLjKzRUBD4G7Ca/4p4dhvrkBcrgwUnThzzqWIussuMLPYa1bOVQdes3EOiJpYfi6phqReQF9gStJxOVdV+BXbzgU/IzQXNSU0a51vZu8kG5JzVYc3oznnnIudN6M555yLnTejFaFZs2bWunXrpMNwzrlK5e233/7GzJoX9ZwnmyK0bt2avLy8pMNwzrlKRdKnxT3nzWjOOedi58nGOedc7DzZOOeci50nG+ecc7HzZOOccy52nmycc87FzpONc8652HmycS7FpEnw5JOwfv22l3XOlZ5f1Olc5PHHoX//cL9pUzjlFDjtNNh/f1Cc84Q6Vw3EWrOR1EvSB5IWShpazDInS5ofzRU+MaX8RknvRbdTUsrHSpojaa6kRyXVj8ovjbYzV9KMaPa+gnV2kzRd0vvRMq3jO2pXGS1ZAmefDbm5MHUqHHUUjBsHBx4IbdvCiBGwcGHSUTpXiZlZLDegJvAxYfrYOsAcoH2hZdoC7wCNo8c7Rn97A88Tal71gLeAhtFzDVPWvxUYGt0/DMiJ7p8PTE5Z7iXgqOh+/YLlirt17drVXPWxYYPZQQeZ1a9v9tFHP5WvXGk2frzZEUeYSWZgtv/+ZnfeabZsWWLhOpe1gDwr5ns1zppNd2ChmS2yMGf6JMKEVKkGA6PN7FsAM/s6Km8PvGJm+Wb2AzAX6BUtswpAkghT4lpU/qKZrYnWfxNoGS3XHqhlZs9Hy32fspxzXHMNvPoq/P3vsMceP5U3bAgDB8ILL8Bnn8FNN8GaNTBkCOy8M/TpAw8/DGvXJha6K+SLL2DaNPjPf2DBgvB47VrwmVSSF+c5mxbAZymPlwL7FVqmHYCk1wg1oRFm9hyhFjRc0i1ADqHWMr9gJUnjgWOjsj8Use+zgWdT9vGdpMeBNsALhNrQxtQVJJ0LnAuw2267lfVYXSX18stw7bVwxhnh/ExxWrSAyy8Pt7lz4cEHYcKE0JmgYUM48cSw/qGHQg3vdpOIZcugQwf45putn6tdGxo1gh12+Olv6v3Uv4ceCj7oe/rFNnmapBOBXmZ2TvT4dGA/MxuSssxTwAbgZEJN5BVgXzP7TtIw4CRgGfA18JaZ3Zaybk3gjqh8fEr5acAQ4FAz+zGKYyzQGVgCTAaeMbOxxcWem5trPupz1bd8OXTsCDk58Pbb0KBB2dbfuDEkqwcegMceg9WroWXLUEM67rh4YnbFGzAAHnkEJk+G7baDlSvhu++2/Ftc2fff/7SdRo3g2WdDxxBXNpLeNrPcIp8srn2tojfgAGBayuMrgSsLLfMP4KyUxzOAbkVsayJwbBHlhwBPpTw+Enif6NxPVLY/8HLK49MJTXd+zqYa27TJrE8fs9q1zfLyKr69H34wmzTJrH17s512Cud7XOY89VQ4pzZ8ePnW37DBbPlys1mzzPbYw6xePbMZM9IaYrVAQuds3gLaSmojqQ5wKjC10DJTgJ4AkpoRmrwWSaopqWlU3gHoAExXsEdULqAPsCB63Bn4J9DHfjr3UxBHI0kFE/ocTkqTnKueRo8Ovc5uvBG6dq349nJyQlfpe++Fr76CG26o+DZd6axaBeedB3vvDVdeWb5t1KoFTZpA584wcya0aQPHHgv/+ld6Y63WistC6bgRzqt8SOiVNiwqu5qQEABE6FE2H3gXODUqrxuVzSec7O8UldcAXouWfQ+YwE+91F4AvgJmR7epKXEcRehk8C5wL1CnpLi9ZlO1zZ5ttt12ZsceG2o46XbGGWZ16pgtWpT+bbutXXBB6C34xhvp2+by5Wbdu5vVrGn24IPp225VRwk1m9jO2VRmfs6m6vrhh3AtzXffwZw5sOOO6d/H559Du3bhl/Ejj6R/++4nM2fCIYfAxRfD3/6W3m2vXg2//jW8+GKoCZ9/fnq3XxWVdM7G+824auXii+GDD0JvsjgSDYSea0OHwqOPhi9DF4916+Ccc0LPsWuvTf/2GzSAp58OnT0uuAD++tf076M68WTjqo3Jk+Gee+CKK+CII+Ld1x/+ALvuGpLbpk3x7qu6uuYa+PBDGDMG6tWLZx9164aehr/5TTgfNHSoX7NTXp5sXLXwySdw7rmhO+vVV8e/v5yc8Et41iy4//7491fdzJ4dOncMHBiGFopT7dqhe/t554V9/u53/gOiPPycTRH8nE3VsmFDaNefPz98SbVpk5n9moWx1RYvho8+gvr1M7Pfqi4/H/bbL5wbmz8/9CLLBLNQu7nxxnBNz/jxIRG5n/g5G1etDR8Ob74ZmlsylWggjBT9t7/Bl196e3863XprqDHeeWfmEg2E9/Ovfw3d2idMCKNGrFuXuf1Xdp5sXJU2Y0b4gjj77HAdTKbtv3/4FTxyJHz6aeb3X9V89FH48XD88XDCCcnEMHToT9dp9e695egDrniebFyVtWwZnH467Lkn3H57cnHccEMYL+2KK5KLoSrYtAkGDw5D0dx5Z7JzDF1wQTiP8/LLcOSRsGJFcrFUFp5sXJW0aVM4ebxiRZh9M67eSqWx667wxz+G3nCvvZZcHJXdPfeEL/dbboFddkk6mjDw6mOPwTvvQM+eobnUFc+TjauSRo2CZ54JzVcdOyYdTRgtukUL7wpdXp9/Hl7Dww+HQYOSjuYnffuGz9miRXDwwd5UWhJPNq7KmTUr1CT69AndVLNBvXrh3FFeXrig1JWeWbh6f8OG0Mkj26boPuKIMOfRN9/AQQeFeXTc1jzZuCrl++/h1FPD6ADjxmXXF9NvfgPdu4fusz/8kHQ0lcfDD4d5g665Bn7+86SjKdr++4cmvoJu9rNnJx1R9vFk46qM/Hz47W9h4cLQNbVp06Qj2lKNGqEr9P/+F2b9dNu2fDlceCF06wYXXZR0NCXr0CEMT1S3bjiH88YbSUeUXTzZuCrh449DE8bEiWGEgEMPTTqioh14YKh53XQTLFmSdDTZ75JL4NtvQ+eAWnHOK5wmbduGKcZ33DGMbDBjRtIRZQ9PNq5SMwtzyHTqFAbYnDQJ/vSnpKMqWcEFnuWde6W6eO650L34yitDraGy2G23UMPZffcw8vfUwrN4VVOebFyltWJFuFDzrLPCBGhz5iRz4WZZtWoFl10WamHe1FK01atDk+gvfgHDhiUdTdnttBO89FL4EdSvX3ivqztPNq5SevHF0KX5iSfCRZMzZoRflJXFFVfAzjuHZiLvCr21YcPgs89C89l22yUdTfk0aRJ6qR18cLgmZ8yYpCNKlicbV6msX//TFAE5OaFmMHQo1KyZdGRlU79+SJL/+Q889FDS0WSX118PIwQMGRLOcVVmDRqE63COPTbU1G65JemIkuOjPhfBR33OTgsWhHHGZs0K0wXcemuyIwNU1KZNoSv0l1+G802V+VjSZd066NwZ1qyBefOqzkjZ69eHoZMefhiuugpGjMiubvnp4qM+u0rNDP75T+jSJVyh/cQT4XFl/3KuUQNuuy1cHT9yZNLRJO/HH8PFuAsWhCanqpJoAOrUCedtBg0KvSUvvbT6TcLmycZltWXLwjzw550XujbPnRseVxUHHQQnnxzmSFm6NOlokrFhA9x9N7RrB3fcEZqbfvnLpKNKv5o1w3FedFH4kTF4MGzcmHRUmRNrspHUS9IHkhZKGlrMMidLmi9pnqSJKeU3Snovup2SUj5W0hxJcyU9Kql+VH5ptJ25kmZIalVoPw0lLZV0Z1zH69Jr2rTQ5fW550KT2XPPZccAjOl2442hSe3//i/pSDIrPz9MQLbnnqFZdOedw3v+978nHVl8Ci7sveoqGDs2NAuvX590VBliZrHcgJrAx8DuQB1gDtC+0DJtgXeAxtHjHaO/vYHngVpAPeAtoGH0XMOU9W8Fhkb3DwNyovvnA5ML7et2YCJw57Zi79q1q7nkrF1rdtFFZmDWvr3Z7NlJRxS/K68Mx/uf/yQdSfzy880eeMBsjz3CMXftavb002abNiUdWWbdfHM4/t69zdasSTqa9ADyrJjv1ThrNt2BhWa2yMzWA5OAvoWWGQyMNrNvAczs66i8PfCKmeWb2Q/AXKBXtMwqAEkCtgcsKn/RzNZE678JtCzYiaSuwE7A9LQfpUurd98NQ5PcfnvojZSXlx2jNsftyivDtRkXX1x12/I3bQoX3e69dzhZXq8e/Otf8NZbobdWVTxhXpLLLoN//OOn3mqrVycdUbziTDYtgM9SHi+NylK1A9pJek3Sm5J6ReVzgF6SciQ1I9Radi1YSdJ44EtgL+COIvZ9NvBstGwN4BbgspKClXSupDxJecuWLSvtMbo0+ve/w4CGX38NTz8d2u+33z7pqDKjQQO4/vrQlbuqjQq9aRM8+mhoEu3fH2rXDvPAzJoVRuaubkkm1W9/G97vmTOr/iRsSXcQqEVoSusJ9AfultTIzKYDzwCvAw8BbwCbT6WZ2VnALsD7wBbXjEs6DcgFbo6KLgCeMbMST7+a2RgzyzWz3ObNm6fh0FxZTJsWptjdffcwYu6xxyYdUeadeSYccED4AnrrraSjqTgzmDIldGU+6aSQdCZPDiM99OsXzl+4MBr4Y4+Fz33PnvDVV8nFsnJlmCohDnG+3Z+TUhshNGt9XmiZpcBUM9tgZp8AHxKSD2Z2nZl1MrOjAEXPbWZmGwlNc5tnIpd0JDAM6GNmP0bFBwBDJC0GRgJnSPpreg7RpcOTT4ZfuHvtFUYG2HnnpCNKRs2aoVv3z34Gxx0XJuSqjMxCzTQ3F44/HtauDaNwv/tu6HnnSWZrffuG1+zjj8Mgsl98kfkYVq4MvQB79Yqpl1xxJ3MqeiPUWhYBbfipg8DehZbpBdwX3W9GaHZrSuhc0DQq7wC8F21PwB5RuQjJY2T0uDOhQ0LbEmIaiHcQyCqPPmpWq5ZZt25mK1YkHU12WLDArEkTsz33NPvmm6SjKZuXXjLr3j2c+N59d7N77zXbsCHpqCqPmTPN6tc3a9fO7H//y9x+v/02vG+1apk9/nj5t0MSHQTMLB8YAkwjNHc9bGbzJF0tqU+02DRguaT5wIvA5Wa2HKgNzIzKxwCnRdsTcJ+kd4F3gZ2Bq6Nt3QzUBx6RNFuSj7Wa5R56KAyc2b07PP88NG6cdETZYc89w4nzxYvDNUXr1iUdUelMnhyGEfrqqzCm2YIFoWmwMkwNkC0OOgiefTbMedSzZ2ZqON99B0cfDe+8E86tHX98TDsqLgtV55vXbOJ3771mNWqYHXKI2apVSUeTnR5+ONQQTj7ZbOPGpKMp2QMPhPfz4IP9/UyHTNVwvv02tCrUrm32r39VfHsk1PXZuSLdc0+YFuCww0K3zwYNko4oO510UhjG5uGHw+Cj2eree+GMM8K5hmef9fczHTJRw/n22zDB2+zZoYNCnz7bXqciPNm4jBo9OgzT0atX6BhQ2cc3i9ull4brjUaODCMhZ5u77w4/HI46Cp56yt/PdIoz4RQkmrlz4fHH4Ve/St+2i+PJxmXMrbeGL86+fUOvq+pyDU1FSGEcrT59wpha//pX0hH95K67wjAzxx4b4srJSTqiqieOhLNiRbim5913Q6I57riKb7M0PNm4jLjhBvjDH+DEE+GRRyrvhFhJqFkzdKbIzQ0XRf73v0lHFBLg734XkuDjj0PduklHVHWlM+EUJJr33gs/+Hr3TluY2+TJxsXKDP7ylzDI5G9+E740a9dOOqrKJycnNDvuvHPy1+DcfHOYYbRfP//hkCnpSDjLl4fegvPnh4ttM33htCcbFxuzML3viBEwcCDcf793g62IHXcMXzgbN8Ixx4Qvj0y7/vow58wpp4RxzurUyXwM1VVFEs7y5aFG8/77IdEcc0xsYRbLk42LhVkYaPCGG0K7/tixlW/q5mzUrh1MnRomkevbN3PX4JiFHw3DhsFpp4XxvLyGmnnlSTjffBNqNO+/H86t9eq17XXi4MnGpd2mTXDhhaFDwIUXhpFtfYiS9OnRI3zZv/566HK8aVO8+zODP/0pNIcOHBi6OnsNNTllSTgFieaDD8KPlCQnpfOvAJdWmzaFgSRHjw41m9tvr96j+sblxBNDd+hHHgnNWnExC9f4XH996LLuNdTsUJqEs2wZHH44fPhhSDRHH53xMLdU3NWe1fnmIwiUT36+2Zlnhqvehw2rfpNhZdqmTWYXXhhe71Gj4tl+wSR2v/td9o9iUB0VN9LA11+b7buvWd26Zs8/n7l4KGEEgcS/2LPx5smm7NavD8OqgNnVVycdTfWRn2/Wt6+ZZDZlSvq2u3Gj2fnnh/fz4ov9h0M2K5xwvvrKbJ99zLbf3uyFFzIbS0nJxpvRXIWtWwcnnBCGVRk5Ev7856Qjqj5q1oSJE8Pspv37w3/+U/FtFjSF/v3vcPnl4dybN4Vmr8JNaocfHqYqeOqpcL4mW/hpPlchP/wQRiZ+4YVwRfn55ycdUfVTcA3OAQeEYUdGj4ZGjUL59tuHv6m3unWL77CxcSOcc07oBDBsGFxzjSeayqAg4RxzTHgPn346jD2YTTzZuHJbtSpcgfz66+HL6cwzk46o+iq4BqdHjzBB2bbUrftT8klNSGvWhJk0//IXuOqq+ON26XPQQZCXF5JN+/ZJR7M1TzauXFasCN0oZ88OF/eddFLSEbl27WDhwjC6wJo1P93Wrt3ycUllNWqEHoS//33SR+PKY889k46geJ5sXJl99VUYMfbDD8P4SpkayM9t2w47QOfOSUfh3NY82bgyWbo0nHRcujScgDzyyKQjcs5VBp5sXKktWhQSzYoVMG1aaCN2zrnS8GTjSmXBglCLWbsWZswIw90751xpebJx2zR3bkg0Erz0Euy7b9IROecqG7+o05Xov/8NF4pttx3MnOmJxjlXPrEmG0m9JH0gaaGkocUsc7Kk+ZLmSZqYUn6jpPei2ykp5WMlzZE0V9KjkupH5ZdG25kraYakVlF5J0lvRNufm7otV7KZM0ONplGjcL9du6Qjcs5VVrElG0k1gdHAMUB7oL+k9oWWaQtcCfQws72Bi6Py3kAXoBOwH3CZpIbRapeYWUcz6wAsAYZE5e8AuVH5o8BNUfka4Ixo+72A2yQ1iuOYq5Lnnw/X0bRoERJN69ZJR+Scq8zirNl0Bxaa2SIzWw9MAvoWWmYwMNrMvgUws6+j8vbAK2aWb2Y/AHMJiQIzWwUgScD2gEXlL5rZmmj9N4GWUfmHZvZRdP9/wNdA8xiOt8qYOjVcO9O2Lbz8ckg4zjlXEXEmmxbAZymPl0ZlqdoB7SS9JulNSQVzyM0BeknKkdQMOAzYtWAlSeOBL4G9gDuK2PfZwLOFCyV1B+oAHxfx3LmS8iTlLVu2rLTHWOVMnhwG1ezYEV58MQyD4pxzFZV0B4FaQFugJ9AfuFtSIzObDjwDvA48BLwBbCxYyczOAnYB3ge2OAcj6TQgF7i5UPnOwAPAWWa21dyGZjbGzHLNLLd58+pZ8XngAfjNb8KAji+8AE2aJB2Rc66qiDPZfE5KbYTQrPV5oWWWAlPNbIOZfQJ8SEg+mNl1ZtbJzI4CFD23mZltJDTNnVBQJulIYBjQx8x+TClvCDwNDDOzN9N0fFVKwUCaPXuGAR0bNtzWGs45V3pxJpu3gLaS2kiqA5wKTC20zBRCrYaouawdsEhSTUlNo/IOQAdguoI9onIBfYAF0ePOwD8Jiabg3A/Rvp8A7jezR+M62Mrsnntg0KAw3tlTT0G9eklH5JyramK7qNPM8iUNAaYBNYFxZjZP0tWE2dymRs8dLWk+oZnscjNbLqkuMDPkE1YBp0XbqwHcF9VURDi3UzCDys1AfeCRaL0lZtYHOBk4BGgqaWC07EAzmx3XsVcm//hHmIOmV68wqGbduklH5JyrihRm8nSpcnNzLS8vL+kwYnfnnXDhhWFOmsceCxduOudceUl628yKHMwq6Q4CLiG33RYSTd++nmicc/HzZFMN3XILXHIJ9OsHDz/sicY5Fz9PNtXMX/8Kl10Wpg6eNAnq1Ek6IudcdeDJphq59lq48kro3x8mTIDatZOOyDlXXXiyqQbMYMQI+POf4fTTw8WbtXxyCedcBnmyqeLM4Kqr4C9/gYEDYfx4qFkz6aicc9WN/76twsxCs9mNN8I558A//wk1/OeFcy4B/tVTRZnB5ZeHRHPeeZ5onHPJ8ppNFWQWujbffjsMGQKjRoUpnZ1zLimebKoYs3Cx5ujRcNFF8Le/eaJxziXPG1aqmCFDQqK57DJPNM657OHJpgp55x24665Qo7npJk80zrns4cmmChk7Ngw9M3y4JxrnXHbxZFNFrFsXRgXo1w8aN046Guec25J3EKgipkyB776Ds89OOhLnymfDhg0sXbqUdevWJR2K24a6devSsmVLapdhzCtPNlXE2LHQujUcdljSkThXPkuXLqVBgwa0bt0aeTtw1jIzli9fztKlS2nTpk2p1/NmtCpg8WKYMQPOOssv3HSV17p162jatKknmiwniaZNm5a5BupfTVXAvfeGvwMHJhmFcxXniaZyKM/75Mmmktu0KQyuedRRsNtuSUfjXOW1fPlyOnXqRKdOnfjZz35GixYtNj9ev359ievm5eXx+9//fpv7OPDAA9MS60svvcRxxx2Xlm1liiebSm7GDFiyBAYNSjoS5zJrwoRwnrJGjfB3woSKba9p06bMnj2b2bNnc95553HJJZdsflynTh3y8/OLXTc3N5dRo0Ztcx+vv/56xYKsxGJNNpJ6SfpA0kJJQ4tZ5mRJ8yXNkzQxpfxGSe9Ft1NSysdKmiNprqRHJdWPyi+NtjNX0gxJrVLWOVPSR9HtzDiPOdPGjQtdnfv2TToS5zJnwgQ491z49NMwRNOnn4bHFU04hQ0cOJDzzjuP/fbbjz/+8Y/897//5YADDqBz584ceOCBfPDBB8CWNY0RI0YwaNAgevbsye67775FEqpfv/7m5Xv27MmJJ57IXnvtxYABAzAzAJ555qvYDtsAAB+YSURBVBn22msvunbtyu9///tt1mBWrFjBr3/9azp06MD+++/P3LlzAXj55Zc318w6d+7M6tWr+eKLLzjkkEPo1KkT++yzDzNnzkzvC1YSM4vlBtQEPgZ2B+oAc4D2hZZpC7wDNI4e7xj97Q08T+gtVw94C2gYPdcwZf1bgaHR/cOAnOj++cDk6H4TYFH0t3F0v3FJsXft2tUqg+XLzbbbzuzCC5OOxLmKmz9/fqmXbdXKLKSZLW+tWqUnluHDh9vNN99sZ555pvXu3dvy8/PNzGzlypW2YcMGMzN7/vnnrV+/fmZm9uKLL1rv3r03r3vAAQfYunXrbNmyZdakSRNbv369mZnVq1dv8/INGza0zz77zDZu3Gj777+/zZw509auXWstW7a0RYsWmZnZqaeeunm7qVL3N2TIEBsxYoSZmc2YMcM6duxoZmbHHXecvfrqq2Zmtnr1atuwYYONHDnSrr32WjMzy8/Pt1WrVpX7NSrq/QLyrJjv1VLVbCTVk1Qjut9OUh9J2+pg3R1YaGaLzGw9MAko/Pt7MDDazL6NEt/XUXl74BUzyzezH4C5QK9omVVRHAK2Bywqf9HM1kTrvwm0jO7/EnjezFZE+3m+YFuV3cSJ8OOP3oTmqp8lS8pWXhEnnXQSNaMZB1euXMlJJ53EPvvswyWXXMK8efOKXKd3795st912NGvWjB133JGvvvpqq2W6d+9Oy5YtqVGjBp06dWLx4sUsWLCA3XfffXOX4v79+28zvldffZXTTz8dgMMPP5zly5ezatUqevTowaWXXsqoUaP47rvvqFWrFt26dWP8+PGMGDGCd999lwYNGpT3ZSmz0jajvQLUldQCmA6cDty7jXVaAJ+lPF4alaVqB7ST9JqkNyUVJIE5QC9JOZKaEWotuxasJGk88CWwF3BHEfs+G3i2DHEg6VxJeZLyli1bto1Dyw7jxkGXLtCpU9KROJdZxXWGiaOTTL169Tbf//Of/8xhhx3Ge++9x5NPPlls99/ttttu8/2aNWsWeb6nNMtUxNChQ7nnnntYu3YtPXr0YMGCBRxyyCG88sortGjRgoEDB3L//fendZ8lKW2yUVRr6AfcZWYnAXunYf+1CE1pPYH+wN2SGpnZdOAZ4HXgIeANYGPBSmZ2FrAL8D5wSuoGJZ0G5AI3lyUQMxtjZrlmltu8efNyH1CmvPNOuHmtxlVH110HOTlbluXkhPI4rVy5khYtwm/VewuuOUijPffck0WLFrF48WIAJk+evM11Dj74YCZEJ6teeuklmjVrRsOGDfn444/Zd999ueKKK+jWrRsLFizg008/ZaeddmLw4MGcc845zJo1K+3HUJxSJxtJBwADgKejsm3NZP85KbURQrPW54WWWQpMNbMNZvYJ8CEh+WBm15lZJzM7ClD03GZmtpHQNHdCSpBHAsOAPmb2YxniqHTGjQuDbv7mN0lH4lzmDRgAY8ZAq1Zh0NlWrcLjAQPi3e8f//hHrrzySjp37pz2mgjA9ttvz1133UWvXr3o2rUrDRo0YIcddihxnREjRvD222/ToUMHhg4dyn333QfAbbfdxj777EOHDh2oXbs2xxxzDC+99BIdO3akc+fOTJ48mYsuuijtx1Cs4k7mpN6AQ4GpwBXR492BUdtYpxbhZHwbfuogsHehZXoB90X3mxGau5oSElnTqLwD8F60PQF7ROUCRgIjo8edCR0S2hbaRxPgE0LngMbR/SYlxZ7tHQTWrjVr3Nisf/+kI3EufcrSQaAqW716tZmZbdq0yc4//3y79dZbE46oaGXtIFCqsdHM7GXgZYCoo8A3ZlbiFUxmli9pCDAtSh7jzGyepKujgKZGzx0taT6hmexyM1suqS4wM7pKdRVwWrS9GsB9khpGyWYOoecZhGaz+sAj0XpLzKyPma2QdA2hRxvA1Wa2ojTHna2mTIFvv/UmNOeqorvvvpv77ruP9evX07lzZ377298mHVJayKK+3SUuFK5/OY+QEN4CGgK3m1mZzotUFrm5uZaXl5d0GMU6+mj48ENYtMjHQnNVx/vvv88vfvGLpMNwpVTU+yXpbTPLLWr50n5VtbfQ5fjXhF5ebQg90lyGffopvPCCD7rpnKtcSvt1VTu6rubXRCf0ia5vcZnlg2465yqj0iabfwKLCVfzvxINBbMqrqBc0QoG3TzyyND7xjnnKotSJRszG2VmLczs2KjTwaeECy1dBv3736EZzTsGOOcqm9IOV7ODpFsLrrCXdAuhluMyaOzYMOjmr3+ddCTOVT2HHXYY06ZN26Lstttu4/zzzy9mDejZsycFnYmOPfZYvvvuu62WGTFiBCNHjixx31OmTGH+/PmbH1911VW88MILZQm/SNk0FUFpm9HGAauBk6PbKmB8XEG5ra1YAU88ES5aq1s36Wicq3r69+/PpEmTtiibNGlSqcYngzBac6NGjcq178LJ5uqrr+bII48s17ayVWmTzc/NbLiFQTUXmdlfCBd2ugx56KEw6ObZZycdiXNV04knnsjTTz+9eaK0xYsX87///Y+DDz6Y888/n9zcXPbee2+GDx9e5PqtW7fmm2++AeC6666jXbt2HHTQQZunIYBwDU23bt3o2LEjJ5xwAmvWrOH1119n6tSpXH755XTq1ImPP/6YgQMH8uijjwIwY8YMOnfuzL777sugQYP48ccfN+9v+PDhdOnShX333ZcFCxaUeHxJT0VQqos6gbWSDjKzVwEk9QDWVnjvrtTGjoXOnX3QTVc9XHwxzJ6d3m126gS33Vb8802aNKF79+48++yz9O3bl0mTJnHyyScjieuuu44mTZqwceNGjjjiCObOnUuHDh2K3M7bb7/NpEmTmD17Nvn5+XTp0oWuXbsC0K9fPwYPHgzAn/70J8aOHcuFF15Inz59OO644zjxxBO32Na6desYOHAgM2bMoF27dpxxxhn8/e9/5+KLLwagWbNmzJo1i7vuuouRI0dyzz33FHt8w4cPp3PnzkyZMoV///vfnHHGGcyePZuRI0cyevRoevTowffff0/dunUZM2YMv/zlLxk2bBgbN25kzZo1xW63tEpbszkPGC1psaTFwJ1A1bistRLwQTedy4zUprTUJrSHH36YLl260LlzZ+bNm7dFk1dhM2fO5PjjjycnJ4eGDRvSp0+fzc+99957HHzwwey7775MmDCh2CkKCnzwwQe0adOGdu3aAXDmmWfyyiuvbH6+X79+AHTt2nXz4J3FSXoqgtIOVzMH6BgNE4OZrZJ0MWGeGRczH3TTVTcl1UDi1LdvXy655BJmzZrFmjVr6Nq1K5988gkjR47krbfeonHjxgwcOLDYqQW2ZeDAgUyZMoWOHTty77338tJLL1Uo3oJpCioyRcHQoUPp3bs3zzzzDD169GDatGmbpyJ4+umnGThwIJdeeilnnHFGhWIt0zXoZrYqGkkA4NIK7dmVyrp1Yarb44+HJk2Sjsa5qq1+/focdthhDBo0aHOtZtWqVdSrV48ddtiBr776imeffbbEbRxyyCFMmTKFtWvXsnr1ap588snNz61evZqdd96ZDRs2bJ4WAKBBgwasXr16q23tueeeLF68mIULFwLwwAMPcOihh5br2JKeiqC052yKogrv3W1TwaCb3jHAuczo378/xx9//ObmtIIh+ffaay923XVXevToUeL6Xbp04ZRTTqFjx47suOOOdOvWbfNz11xzDfvttx/Nmzdnv/3225xgTj31VAYPHsyoUaM2dwwAqFu3LuPHj+ekk04iPz+fbt26cd5555XruEaMGMGgQYPo0KEDOTk5W0xF8OKLL1KjRg323ntvjjnmGCZNmsTNN99M7dq1qV+/flomWSvVQJxFrigtMbMY5sVLXjYNxOmDbrrqwgfirFzKOhBniTUbSaspegw0AduXN0hXOgWDbg4f7onGOVe5lZhszKziXRBcufmgm865qsJ/L2epgkE3jzjCB910zlV+nmyyVMGgm94xwFUn5T2H7DKrPO+TJ5ssNW6cD7rpqpe6deuyfPlyTzhZzsxYvnw5dcs4SGNFuj67mHz7LTz+OAwe7INuuuqjZcuWLF26lGXLliUdituGunXr0rJlyzKt48kmC02cGAbd9OFpXHVSu3Zt2rRpk3QYLiaxNqNJ6iXpA0kLJQ0tZpmTJc2XNE/SxJTyGyW9F91OSSkfK2mOpLmSHpVUPyo/RNIsSfmSTiy0j5ui7b8vaZSkrL4gddy4MOhm585JR+Kcc+kRW7KRVBMYDRwDtAf6S2pfaJm2wJVADzPbG7g4Ku8NdAE6AfsBlxWMywZcYmYdzawDsAQYEpUvAQYCE0kh6UCgB9AB2AfoBpRvvIcMmD0bZs3yWo1zrmqJs2bTHVgYzX+zHpgE9C20zGBgtJl9C2BmX0fl7YFXzCzfzH4gDPjZK1pmFUBUO9me6KJTM1tsZnOBTYX2YUBdoA6wHVAb+CqdB5pOPuimc64qijPZtAA+S3m8NCpL1Q5oJ+k1SW9K6hWVzwF6ScqR1Aw4DNi1YCVJ44Evgb2AO0oKwszeAF4Evohu08zs/fIfVnzWroUHH/RBN51zVU/SXZ9rAW2BnkB/4G5JjcxsOvAM8DrwEPAGsLFgJTM7C9gFeB84hRJI2gP4BdCSkOwOl3RwEcudKylPUl5SvWHGjQs90UqY8tw55yqlOJPN56TURghf9p8XWmYpMNXMNpjZJ8CHhOSDmV1nZp3M7CjCWGwfpq5oZhsJTXMnbCOO44E3zex7M/seeBY4oPBCZjbGzHLNLLd58+alPsh02bABbroJevSAg7dKhc45V7nFmWzeAtpKaiOpDnAqMLXQMlMItRqi5rJ2wCJJNSU1jco7EE7uT1ewR1QuoA9Q8sTboePAoZJqSapN6ByQdc1oEyfCkiXwf/8H2d1Xzjnnyi6262zMLF/SEGAaUBMYZ2bzJF0N5JnZ1Oi5oyXNJzSTXW5myyXVBWZGPZRXAadF26sB3Bf1TBPh3M75AJK6AU8AjYFfSfpL1MPtUeBw4F1CZ4HnzOyn2YyywMaNcMMN0LEjHHNM0tE451z6lXs+m6os0/PZPPYYnHgiTJ4MJ5+csd0651xalTSfTdIdBKo9M7j+emjbFk7Y1tkn55yrpHy4moRNnx4u4hw7FmrWTDoa55yLh9dsEnb99dCyJZx2WtKROOdcfLxmk6BXX4VXXoHbb4c6dZKOxjnn4uM1mwTdcAM0awbnnJN0JM45Fy9PNgmZPRueeQYuuQRycpKOxjnn4uXJJiE33AANGsAFFyQdiXPOxc+TTQI+/BAeeQR+9zto1CjpaJxzLn6ebBJw441hGoGLL046EuecywxPNhm2ZAncf3/oFLDTTklH45xzmeHJJsNuuSX8veyyZONwzrlM8mSTQV9/DXffHS7gbNUq6Wiccy5zPNlk0O23w7p1cMUVSUfinHOZ5ckmQ1auhDvvDINt7rVX0tE451xmebLJkLvuglWr4Mork47EOecyz5NNBqxZA3/7G/TqBV26JB2Nc85lniebDBg7FpYtC1M+O+dcdeTJJmbr18PNN8NBB8HBBycdjXPOJcOnGIjZhAnw2Wfwz38mHYlzziXHazZpNGECtG4NNWqEvw88AH/9K3TqFM7XOOdcdeU1mzSZMAHOPTd0BgD49NMwJM369fDwwyAlG59zziUp1pqNpF6SPpC0UNLQYpY5WdJ8SfMkTUwpv1HSe9HtlJTysZLmSJor6VFJ9aPyQyTNkpQv6cRC+9hN0nRJ70f7ap3uYx027KdEU2D9eqhVC/r1S/fenHOucokt2UiqCYwGjgHaA/0ltS+0TFvgSqCHme0NXByV9wa6AJ2A/YDLJDWMVrvEzDqaWQdgCTAkKl8CDAQmsrX7gZvN7BdAd+DrdB1ngSVLii7Pz4eaNdO9N+ecq1zirNl0Bxaa2SIzWw9MAvoWWmYwMNrMvgUws4Ik0B54xczyzewHYC7QK1pmFYAkAdsDFpUvNrO5wKbUHUQJrpaZPR8t972ZFaqDVNxuu5Wt3DnnqpM4k00L4LOUx0ujslTtgHaSXpP0pqSC0+hzgF6SciQ1Aw4Ddi1YSdJ44EtgL+CObcTRDvhO0uOS3pF0c1Tr2oKkcyXlScpbtmxZWY4TgOuu23p659q14frry7wp55yrcpLujVYLaAv0BPoDd0tqZGbTgWeA14GHgDeAjQUrmdlZwC7A+8AplKwWcDBwGdAN2J3Q3LYFMxtjZrlmltu8efMyH8iAATBmzE+jOdeoAf/4Ryh3zrnqLs5k8zkptRGgZVSWaikw1cw2mNknwIeE5IOZXWdmnczsKEDRc5uZ2UZC09wJ24hjKTA7as7LB6YQzgel3YAB8Pjj4f6118KgQXHsxTnnKp84k81bQFtJbSTVAU4FphZaZgqhVkPUXNYOWCSppqSmUXkHoAMwXcEeUbmAPsCCUsTRSFJBdeVwYH5FD644N9wADRvCBRfEtQfnnKt8YrvOxszyJQ0BpgE1gXFmNk/S1UCemU2Nnjta0nxCM9nlZrZcUl1gZsgnrAJOi7ZXA7gv6pkmwrmd8wEkdQOeABoDv5L0FzPb28w2SroMmBElqLeBu+M45oUL4bHHwsjOO+wQxx6cc65ykpklHUPWyc3Ntby8vDKvZwZPPw3du8OOO8YQmHPOZTFJb5tZblHP+QgCaSTBccclHYVzzmWfpHujOeecqwY82TjnnIudJxvnnHOx82STRQpPUTBhQtIROedcengHgSxR1BQF554b7vsoBM65ys5rNlmiqCkK1qwJ5c45V9l5sskSxU1RUFy5c85VJp5ssoRPUeCcq8o82WSJoqYoyMkJ5c45V9l5sskSqVMUSOHvmDHeOcA5VzV4b7QsMmCAJxfnXNXkNRvnnHOx82TjnHMudp5snHPOxc6TjXPOudh5snHOORc7TzbOOedi58nGOedc7DzZOOeci12syUZSL0kfSFooaWgxy5wsab6keZImppTfKOm96HZKSvlYSXMkzZX0qKT6UfkhkmZJypd0YhH7aShpqaQ74zhW55xzxYst2UiqCYwGjgHaA/0ltS+0TFvgSqCHme0NXByV9wa6AJ2A/YDLJDWMVrvEzDqaWQdgCTAkKl8CDAQmUrRrgFfSc3TOOefKIs6aTXdgoZktMrP1wCSgb6FlBgOjzexbADP7OipvD7xiZvlm9gMwF+gVLbMKQJKA7QGLyheb2VxgU+FAJHUFdgKmp/cQnXPOlUacyaYF8FnK46VRWap2QDtJr0l6U1KvqHwO0EtSjqRmwGHArgUrSRoPfAnsBdxRUhCSagC3AJdtY7lzJeVJylu2bNm2j84551ypJd1BoBbQFugJ9AfultTIzKYDzwCvAw8BbwAbC1Yys7OAXYD3gVMo2QXAM2a2tKSFzGyMmeWaWW7z5s3LeTjOOeeKEmey+ZyU2gjQMipLtRSYamYbzOwT4ENC8sHMrjOzTmZ2FKDouc3MbCOhae6EbcRxADBE0mJgJHCGpL+W75Ccc86VR5zJ5i2graQ2kuoApwJTCy0zhVCrIWouawcsklRTUtOovAPQAZiuYI+oXEAfYEFJQZjZADPbzcxaE5rS7jezInvGVXYTJkDr1lCjRvg7YULSEbnKxD8/Lk6xJRszyyf0FJtGaO562MzmSbpaUp9osWnAcknzgReBy81sOVAbmBmVjwFOi7Yn4D5J7wLvAjsDVwNI6iZpKXAS8E9J8+I6tmw0YQKcey58+imYhb/nnutfGK500vH58WSVrGx//WVmSceQdXJzcy0vLy/pMMqkdevwBVFYq1aweHGmo3GVTUU/PwXJas2an8pycny22UzJltdf0ttmllvkc55stlYZk02NGuEXaWESbNqqM7hzW6ro58d/7CQrW17/kpJN0r3RXJrstlvZyouS7dVwF5+Kfn6WLClbeTaq6Oc/yf+fSvH6m5nfCt26du1qlc2DD5rl5JiF36fhlpMTyjOxfjZ48EGzVq3MpPC3MsWetIq+/61abbluwa1VqzijTp/K/v+TLa8/kGfFfK8m/sWejbfKmGzMKvZlmy0f1vJKxz97RZNV0skuyfiT/rKtqIp+/pP+/8mW19+TTTVJNhUhFf3PIiUdWelU9J+9sv+yTXr/BTFU1mRb0c9/Nvz/JP36m3my8WRTCun4ZZbkL+uK/rNX9l+2Se8/HZKsWWXD+58NyaKiPNl4stmmpH/ZJ/1lUdl/2Sa9/4pK+v1P+vObDTXTdPBk48mmVJI855P0l0XS8VdU0vuvqKR/LJglWzOv7O9fAU82nmxilw01gySbYZL+ZZr0/isq6WbQpFX2mmkBTzaebGJX2WsGZsn3Rkt6/SQlXbNNWlX4/Jt5svFkkwGVvWaQND/+5LuuJynp9z9d+/dk48kmI6rzL/OKyoZftkmrzu+/WbLHn67PX0nJxsdGK0JlHBvNVW4+tp1LUro+fz42mnNZLh1j2zlXXpn4/HmycS4LXHddGBI+VU5OKHcubpn4/HmycS4LDBgQ5h5p1So0XbRq5XPBuMzJxOfPz9kUwc/ZOOdc2fk5G+ecc4nyZOOccy52nmycc87FzpONc8652Hmycc45FzvvjVYEScuAT5OOowTNgG+SDqIEHl/FeHwV4/FVTEXia2VmzYt6wpNNJSQpr7juhdnA46sYj69iPL6KiSs+b0ZzzjkXO082zjnnYufJpnIak3QA2+DxVYzHVzEeX8XEEp+fs3HOORc7r9k455yLnScb55xzsfNkk4Uk7SrpRUnzJc2TdFERy/SUtFLS7Oh2VQJxLpb0brT/rYbJVjBK0kJJcyV1yWBse6a8NrMlrZJ0caFlMvoaShon6WtJ76WUNZH0vKSPor+Ni1n3zGiZjySdmcH4bpa0IHr/npDUqJh1S/wsxBjfCEmfp7yHxxazbi9JH0SfxaEZjG9ySmyLJc0uZt1MvH5Ffq9k7DNY3HzRfkvuBuwMdInuNwA+BNoXWqYn8FTCcS4GmpXw/LHAs4CA/YH/JBRnTeBLwgVnib2GwCFAF+C9lLKbgKHR/aHAjUWs1wRYFP1tHN1vnKH4jgZqRfdvLCq+0nwWYoxvBHBZKd7/j4HdgTrAnML/T3HFV+j5W4CrEnz9ivxeydRn0Gs2WcjMvjCzWdH91cD7QItkoyqXvsD9FrwJNJK0cwJxHAF8bGaJjgphZq8AKwoV9wXui+7fB/y6iFV/CTxvZivM7FvgeaBXJuIzs+lmlh89fBNome79llYxr19pdAcWmtkiM1sPTCK87mlVUnySBJwMPJTu/ZZWCd8rGfkMerLJcpJaA52B/xTx9AGS5kh6VtLeGQ0sMGC6pLclnVvE8y2Az1IeLyWZpHkqxf+TJ/0a7mRmX0T3vwR2KmKZbHkdBxFqqkXZ1mchTkOiZr5xxTQBZcPrdzDwlZl9VMzzGX39Cn2vZOQz6Mkmi0mqDzwGXGxmqwo9PYvQLNQRuAOYkun4gIPMrAtwDPA7SYckEEOJJNUB+gCPFPF0NryGm1lor8jKaxEkDQPygQnFLJLUZ+HvwM+BTsAXhKaqbNSfkms1GXv9SvpeifMz6MkmS0mqTfhATDCzxws/b2arzOz76P4zQG1JzTIZo5l9Hv39GniC0FyR6nNg15THLaOyTDoGmGVmXxV+IhteQ+CrgqbF6O/XRSyT6OsoaSBwHDAg+jLaSik+C7Ews6/MbKOZbQLuLma/Sb9+tYB+wOTilsnU61fM90pGPoOebLJQ1L47FnjfzG4tZpmfRcshqTvhvVyewRjrSWpQcJ9wIvm9QotNBc6IeqXtD6xMqa5nSrG/KJN+DSNTgYKePWcC/ypimWnA0ZIaR81ER0dlsZPUC/gj0MfM1hSzTGk+C3HFl3oO8Phi9vsW0FZSm6imeyrhdc+UI4EFZra0qCcz9fqV8L2Smc9gnL0f/FbuXiMHEaqyc4HZ0e1Y4DzgvGiZIcA8Qs+aN4EDMxzj7tG+50RxDIvKU2MUMJrQE+hdIDfDMdYjJI8dUsoSew0JSe8LYAOhzftsoCkwA/gIeAFoEi2bC9yTsu4gYGF0OyuD8S0ktNUXfA7/ES27C/BMSZ+FDMX3QPTZmkv40ty5cHzR42MJva8+zmR8Ufm9BZ+5lGWTeP2K+17JyGfQh6txzjkXO29Gc845FztPNs4552LnycY551zsPNk455yLnScb55xzsfNk41wGSdqoLUejTtsIxJJap4447Fw2qZV0AM5VM2vNrFPSQTiXaV6zcS4LRPOZ3BTNafJfSXtE5a0l/TsaaHKGpN2i8p0U5peZE90OjDZVU9Ld0Xwl0yVtHy3/+2gek7mSJiV0mK4a82TjXGZtX6gZ7ZSU51aa2b7AncBtUdkdwH1m1oEwCOaoqHwU8LKFQUS7EK48B2gLjDazvYHvgBOi8qFA52g758V1cM4Vx0cQcC6DJH1vZvWLKF8MHG5mi6LBEr80s6aSviEMwbIhKv/CzJpJWga0NLMfU7bRmjDnSNvo8RVAbTO7VtJzwPeEka2nWDQAqXOZ4jUb57KHFXO/LH5Mub+Rn87L9iaMU9cFeCsaidi5jPFk41z2OCXl7xvR/dcJoxQDDABmRvdnAOcDSKopaYfiNiqpBrCrmb0IXAHsAGxVu3IuTv7rxrnM2l7S7JTHz5lZQffnxpLmEmon/aOyC4Hxki4HlgFnReUXAWMknU2owZxPGHG4KDWBB6OEJGCUmX2XtiNyrhT8nI1zWSA6Z5NrZt8kHYtzcfBmNOecc7Hzmo1zzrnYec3GOedc7DzZOOeci50nG+ecc7HzZOOccy52nmycc87F7v8BMkHz6/nyVXoAAAAASUVORK5CYII=\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 295
        },
        "id": "efEYZanpg6kv",
        "outputId": "7fed052d-5e66-4904-b359-f484e0a7bfb9"
      },
      "source": [
        "# accuracy\r\n",
        "\r\n",
        "plt.clf()  # 그래프 초기화\r\n",
        "\r\n",
        "acc = history_dict['accuracy']\r\n",
        "val_acc = history_dict['val_accuracy']\r\n",
        "\r\n",
        "plt.plot(epochs, acc, 'bo', label='Training acc')\r\n",
        "plt.plot(epochs, val_acc, 'b', label='Validation acc')\r\n",
        "plt.title('Training and Validation Accuracy')\r\n",
        "plt.xlabel('Epochs')\r\n",
        "plt.ylabel('Accuracy')\r\n",
        "plt.legend()\r\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "avvMN7RCm2ti"
      },
      "source": [
        "training data에 과대적합되었음을 알 수 있다.  \r\n",
        "\r\n",
        "과대적합을 완화하기 위해, 새로운 신경망을 4번의 에포크 동안만 훈련하고  \r\n",
        "test data로 평가해본다."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "iDdn7WpXoh-8"
      },
      "source": [
        "### 2-2) 신경망 모델 만들기(2)\r\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "6sLWMae2oh_I",
        "outputId": "ff0cebd0-9483-4249-cbab-8d54b986b1b9"
      },
      "source": [
        "mymodel = models.Sequential()\r\n",
        "mymodel.add(layers.Dense(16, activation='relu',input_shape=(10000,)))  # 16개의 은닉층\r\n",
        "mymodel.add(layers.Dense(16, activation='relu'))\r\n",
        "mymodel.add(layers.Dense(1, activation='sigmoid'))\r\n",
        "\r\n",
        "# 모델 컴파일\r\n",
        "mymodel.compile(optimizer='rmsprop',\r\n",
        "                loss='binary_crossentropy',\r\n",
        "                metrics=['accuracy'])\r\n",
        "\r\n",
        "mymodel.fit(xtrain, ytrain, epochs=4, batch_size=512)\r\n",
        "results = mymodel.evaluate(xtest, ytest)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Epoch 1/4\n",
            "49/49 [==============================] - 2s 23ms/step - loss: 0.6932 - accuracy: 0.4991\n",
            "Epoch 2/4\n",
            "49/49 [==============================] - 1s 23ms/step - loss: 0.6932 - accuracy: 0.4986\n",
            "Epoch 3/4\n",
            "49/49 [==============================] - 1s 24ms/step - loss: 0.6932 - accuracy: 0.4983\n",
            "Epoch 4/4\n",
            "49/49 [==============================] - 1s 24ms/step - loss: 0.6932 - accuracy: 0.4971\n",
            "782/782 [==============================] - 1s 2ms/step - loss: 0.6932 - accuracy: 0.5000\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "yuLo_Yy4kcPO",
        "outputId": "baca146e-bab2-409d-dc6b-3ecb6fbd3a7c"
      },
      "source": [
        "results  # 50%의 정확도?"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[0.6931557655334473, 0.4999600052833557]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 69
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ktB4Eq3TpX0F"
      },
      "source": [
        "### 2-5) 훈련된 모델로 새로운 데이터에 대해 예측하기"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "qB5y6yyZplyk",
        "outputId": "4b5e53fa-d949-488f-df29-56de108c6f1d"
      },
      "source": [
        "mymodel.predict(xtest)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([[0.60192627],\n",
              "       [0.49975878],\n",
              "       [0.49975878],\n",
              "       ...,\n",
              "       [0.49975878],\n",
              "       [0.49975878],\n",
              "       [0.49975878]], dtype=float32)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 73
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "YcsqYPzHr72U"
      },
      "source": [
        "0.99 or more, 0.01 or less : 샘플의 분류에 대해 확신을 가지고 있음\r\n",
        "\r\n",
        "0.6, 0,4 ... : 확신 부족"
      ]
    }
  ]
}