chapter04_getting-started-with-neural-networks.ipynb

{
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "zZs3aCABaeul"
      },
      "source": [
        "# Getting started with neural networks: Classification and regression"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ypOu1Kujaeun"
      },
      "source": [
        "## Classifying movie reviews: A binary classification example"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "4NI-IAC4aeuo"
      },
      "source": [
        "### The IMDB dataset"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "50,000 highly polarized review from the Internet movie database.  \n",
        "25000 traing data, 25000 test data  \n",
        "50% positive, 50% negative "
      ],
      "metadata": {
        "id": "QBwequUteKkq"
      }
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "vFK4Gei0aeuo"
      },
      "source": [
        "**Loading the IMDB dataset**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "peErQmUlaeuq",
        "outputId": "c0ad1d00-28c9-495d-89b2-65408692913c",
        "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",
            "17465344/17464789 [==============================] - 0s 0us/step\n",
            "17473536/17464789 [==============================] - 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) # only keep 10,000 most frequently occuriring word in the training data"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "uP9ABT7_aeus",
        "outputId": "543243ec-97e9-4e29-98f1-08eade2d5f7e",
        "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": null,
      "metadata": {
        "id": "FuXsgApTaeuu",
        "outputId": "1db73d52-be7d-4b81-e04e-8efa933c16b5",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "1"
            ]
          },
          "metadata": {},
          "execution_count": 5
        }
      ],
      "source": [
        "train_labels[0]"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "zJs1G8IHaeuv",
        "outputId": "94e7d4b5-c13a-47d4-cf01-a7213a5571ad",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "9999"
            ]
          },
          "metadata": {},
          "execution_count": 6
        }
      ],
      "source": [
        "max([max(sequence) for sequence in train_data]) # because we  restricited to the top 10000 most frequent words"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "yoFmRjhfaeux"
      },
      "source": [
        "**Decoding reviews back to text**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "HwlANfMNaeuy",
        "outputId": "834f2736-59aa-49ce-a08a-bb7fd678c788",
        "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",
            "1646592/1641221 [==============================] - 0s 0us/step\n",
            "1654784/1641221 [==============================] - 0s 0us/step\n"
          ]
        }
      ],
      "source": [
        "word_index = imdb.get_word_index() # word -> integer index 매핑 dictionary\n",
        "reverse_word_index = dict(\n",
        "    [(value, key) for (key, value) in word_index.items()])\n",
        "decoded_review = \" \".join( #.join -> list를 문자열로 \n",
        "    [reverse_word_index.get(i - 3, \"?\") for i in train_data[0]])\n",
        "# indices are offset by 3 because 0,1,2 are \"padding\",\"start of sequence\", \"unknown\""
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "decoded_review"
      ],
      "metadata": {
        "id": "AdfiE7rsg85S",
        "outputId": "6d4b5b98-c6a2-4a48-c7c6-69ed3e543074",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 162
        }
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "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\""
            ],
            "application/vnd.google.colaboratory.intrinsic+json": {
              "type": "string"
            }
          },
          "metadata": {},
          "execution_count": 14
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "rYeOQmN4aeuz"
      },
      "source": [
        "### Preparing the data"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Nb6FTZ7yaeu0"
      },
      "source": [
        "**Encoding the integer sequences via multi-hot encoding**"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "원-핫 인코딩은 단어 집합의 크기를 벡터의 차원으로 하고, 표현하고 싶은 단어의 인덱스에 1의 값을 부여하고, 다른 인덱스에는 0을 부여하는 단어의 벡터 표현 방식입니다. 이렇게 표현된 벡터를 원-핫 벡터(One-Hot vector)라고 합니다."
      ],
      "metadata": {
        "id": "OyC2w6peBVNM"
      }
    },
    {
      "cell_type": "code",
      "execution_count": 11,
      "metadata": {
        "id": "ygFD9vwtaeu1"
      },
      "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": null,
      "metadata": {
        "id": "TDg2Dafpaeu2",
        "outputId": "d1b9a276-4016-457f-fadc-8179693ff6c4",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([0., 1., 1., ..., 0., 0., 0.])"
            ]
          },
          "metadata": {},
          "execution_count": 16
        }
      ],
      "source": [
        "x_train[0]"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "yh1JiG80aeu3"
      },
      "outputs": [],
      "source": [
        "y_train = np.asarray(train_labels).astype(\"float32\")\n",
        "y_test = np.asarray(test_labels).astype(\"float32\")"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "2iqa51yGaeu4"
      },
      "source": [
        "### Building your model"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "2 intermediate layers with 16 units each  \n",
        "3rd layer will output the scalar prediction regarding the sentiment of the curruent revies\n"
      ],
      "metadata": {
        "id": "RVVW2wtwkKTT"
      }
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Q2Xdz5lqaeu4"
      },
      "source": [
        "**Model definition**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 17,
      "metadata": {
        "id": "Yl4rQC9Faeu5"
      },
      "outputs": [],
      "source": [
        "from tensorflow import keras\n",
        "from tensorflow.keras import layers\n",
        "\n",
        "model = keras.Sequential([\n",
        "    layers.Dense(16, activation=\"relu\"), # 중간 계층 -> relu 함수 사용(음수를 0으로 만드는 함수)\n",
        "    layers.Dense(16, activation=\"relu\"), \n",
        "    layers.Dense(1, activation=\"sigmoid\") # 임의의 값을 [0,1] interval로 나타냄 (1을 가질 가능성)\n",
        "])"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "more unit(a higher-dimensional representation spae) -> 더 복잡한 표현을 학습할 수 있으나, overfit 될 수 있다. "
      ],
      "metadata": {
        "id": "-noyG_-BlT67"
      }
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "XTnl00Koaeu6"
      },
      "source": [
        "**Compiling the model**"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "- loss choice : binary classification, output of you model is a probability(single unnit layer with a sigmoid activation)   \n",
        "  -> **binary cossentropy loss** is best chois \n",
        "- optimizer choice : rmsprob is usually good defalt choice "
      ],
      "metadata": {
        "id": "vRPs8WHslwG9"
      }
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "k9VbT2Rcaeu6"
      },
      "outputs": [],
      "source": [
        "model.compile(optimizer=\"rmsprop\",\n",
        "              loss=\"binary_crossentropy\",\n",
        "              metrics=[\"accuracy\"])"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "szk9qgmzaeu7"
      },
      "source": [
        "### Validating your approach"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Uubjpl8qaeu8"
      },
      "source": [
        "**Setting aside a validation set**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "959tolXHaeu8"
      },
      "outputs": [],
      "source": [
        "x_val = x_train[:10000] # train data 중 10000개를 validation data에 할당 \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": "1AlvnrChaeu9"
      },
      "source": [
        "**Training your model**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "MZW4j_cWaeu9",
        "outputId": "c7dc45da-0fd4-4d4e-fea4-0521bd0953e1",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/20\n",
            "30/30 [==============================] - 2s 55ms/step - loss: 0.5268 - accuracy: 0.7797 - val_loss: 0.3998 - val_accuracy: 0.8694\n",
            "Epoch 2/20\n",
            "30/30 [==============================] - 1s 41ms/step - loss: 0.3224 - accuracy: 0.8985 - val_loss: 0.3151 - val_accuracy: 0.8852\n",
            "Epoch 3/20\n",
            "30/30 [==============================] - 1s 40ms/step - loss: 0.2369 - accuracy: 0.9247 - val_loss: 0.2839 - val_accuracy: 0.8922\n",
            "Epoch 4/20\n",
            "30/30 [==============================] - 1s 41ms/step - loss: 0.1865 - accuracy: 0.9424 - val_loss: 0.2769 - val_accuracy: 0.8882\n",
            "Epoch 5/20\n",
            "30/30 [==============================] - 1s 41ms/step - loss: 0.1521 - accuracy: 0.9521 - val_loss: 0.2949 - val_accuracy: 0.8835\n",
            "Epoch 6/20\n",
            "30/30 [==============================] - 1s 41ms/step - loss: 0.1254 - accuracy: 0.9623 - val_loss: 0.2923 - val_accuracy: 0.8837\n",
            "Epoch 7/20\n",
            "30/30 [==============================] - 1s 41ms/step - loss: 0.1049 - accuracy: 0.9694 - val_loss: 0.3030 - val_accuracy: 0.8837\n",
            "Epoch 8/20\n",
            "30/30 [==============================] - 1s 40ms/step - loss: 0.0863 - accuracy: 0.9755 - val_loss: 0.3246 - val_accuracy: 0.8798\n",
            "Epoch 9/20\n",
            "30/30 [==============================] - 1s 41ms/step - loss: 0.0715 - accuracy: 0.9801 - val_loss: 0.3427 - val_accuracy: 0.8796\n",
            "Epoch 10/20\n",
            "30/30 [==============================] - 1s 40ms/step - loss: 0.0597 - accuracy: 0.9857 - val_loss: 0.3718 - val_accuracy: 0.8757\n",
            "Epoch 11/20\n",
            "30/30 [==============================] - 1s 41ms/step - loss: 0.0480 - accuracy: 0.9884 - val_loss: 0.3880 - val_accuracy: 0.8783\n",
            "Epoch 12/20\n",
            "30/30 [==============================] - 1s 40ms/step - loss: 0.0392 - accuracy: 0.9909 - val_loss: 0.4327 - val_accuracy: 0.8710\n",
            "Epoch 13/20\n",
            "30/30 [==============================] - 1s 40ms/step - loss: 0.0331 - accuracy: 0.9930 - val_loss: 0.4562 - val_accuracy: 0.8695\n",
            "Epoch 14/20\n",
            "30/30 [==============================] - 1s 40ms/step - loss: 0.0245 - accuracy: 0.9964 - val_loss: 0.4967 - val_accuracy: 0.8669\n",
            "Epoch 15/20\n",
            "30/30 [==============================] - 1s 40ms/step - loss: 0.0219 - accuracy: 0.9962 - val_loss: 0.5073 - val_accuracy: 0.8698\n",
            "Epoch 16/20\n",
            "30/30 [==============================] - 1s 40ms/step - loss: 0.0156 - accuracy: 0.9977 - val_loss: 0.5319 - val_accuracy: 0.8689\n",
            "Epoch 17/20\n",
            "30/30 [==============================] - 1s 41ms/step - loss: 0.0125 - accuracy: 0.9986 - val_loss: 0.5816 - val_accuracy: 0.8673\n",
            "Epoch 18/20\n",
            "30/30 [==============================] - 1s 41ms/step - loss: 0.0109 - accuracy: 0.9987 - val_loss: 0.6007 - val_accuracy: 0.8671\n",
            "Epoch 19/20\n",
            "30/30 [==============================] - 1s 40ms/step - loss: 0.0058 - accuracy: 0.9997 - val_loss: 0.6657 - val_accuracy: 0.8609\n",
            "Epoch 20/20\n",
            "30/30 [==============================] - 1s 40ms/step - loss: 0.0064 - accuracy: 0.9995 - val_loss: 0.6692 - val_accuracy: 0.8665\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": null,
      "metadata": {
        "id": "pVSRWxpGaeu-",
        "outputId": "5444577f-6091-476b-f1a7-d9d911725237",
        "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": 22
        }
      ],
      "source": [
        "history_dict = history.history\n",
        "history_dict.keys()"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "gHGo6R8Iaeu-"
      },
      "source": [
        "**Plotting the training and validation loss**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "ECEaj0LUaeu_",
        "outputId": "c3a8e1a3-3fe7-4f23-c422-c746b0e6b0a0",
        "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": [
        "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": "deFNuYUdaevA"
      },
      "source": [
        "**Plotting the training and validation accuracy**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "0DLproQvaevA",
        "outputId": "9e9fe279-203c-4be5-d661-b59b2e7036b6",
        "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",
      "source": [
        "약 4 epochs 이후로는 loss가 커지고 accuracy가 줄어듦\n",
        "-> 4 epochs까지만 진행 "
      ],
      "metadata": {
        "id": "0qcizxlVoa65"
      }
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "sAuLzVkUaevB"
      },
      "source": [
        "**Retraining a model from scratch**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "DMrdxfpwaevB",
        "outputId": "bb9c9617-2ca0-454c-a550-c01e4c9f9197",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/4\n",
            "49/49 [==============================] - 2s 30ms/step - loss: 0.4843 - accuracy: 0.8181\n",
            "Epoch 2/4\n",
            "49/49 [==============================] - 1s 29ms/step - loss: 0.2736 - accuracy: 0.9048\n",
            "Epoch 3/4\n",
            "49/49 [==============================] - 1s 29ms/step - loss: 0.2065 - accuracy: 0.9256\n",
            "Epoch 4/4\n",
            "49/49 [==============================] - 1s 29ms/step - loss: 0.1703 - accuracy: 0.9403\n",
            "782/782 [==============================] - 2s 2ms/step - loss: 0.2941 - accuracy: 0.8850\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": null,
      "metadata": {
        "id": "8iCX8FkYaevC",
        "outputId": "1b8355ea-8bb9-4fcb-b87c-dd0b014c1f2d",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[0.29411429166793823, 0.8850399851799011]"
            ]
          },
          "metadata": {},
          "execution_count": 26
        }
      ],
      "source": [
        "results #test loss, accuracy"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "BC8F-hufaevC"
      },
      "source": [
        "### Using a trained model to generate predictions on new data"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "_hJ0W3jDaevD",
        "outputId": "17e85b30-8c65-404a-b12d-08bbd63e3090",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([[0.22858968],\n",
              "       [0.9999449 ],\n",
              "       [0.9672936 ],\n",
              "       ...,\n",
              "       [0.161401  ],\n",
              "       [0.09347045],\n",
              "       [0.82443804]], dtype=float32)"
            ]
          },
          "metadata": {},
          "execution_count": 27
        }
      ],
      "source": [
        "model.predict(x_test)\n",
        "\n",
        "# 일부 표본의 경우 모델이 신뢰 가능하지만(0.99 이상, 0.01이하), 다른 표본(0.6,0.4)의 경우 신뢰도가 낮다 "
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "BGf57AduaevD"
      },
      "source": [
        "### Further experiments"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "- 최종 분류 전 2개의 layor 사용 -> 1개 또는 3개의 layor 사용해보기\n",
        "- 더 많거나 적은 units 사용해보기\n",
        "- binary crossentropy 대신 mse사용해보기\n",
        "- relu 대신 tanh 활성화함수 사용해보기 "
      ],
      "metadata": {
        "id": "c-cDbtKSq2_T"
      }
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "r8G_sqjKaevE"
      },
      "source": [
        "### Wrapping up"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "RjAH3wcEaevE"
      },
      "source": [
        "## Classifying newswires: A multiclass classification example"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "- singlelabel multiclass classification : 하나의 data는 한 class에만 속함\n",
        "- multilabel multiclass classification : 하나의 data가 여러 class에 속함"
      ],
      "metadata": {
        "id": "DOXRsgoh_eXJ"
      }
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "tls7jta0aevE"
      },
      "source": [
        "### The Reuters dataset"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "reuters가 출간한 짧은 뉴스 와이어  \n",
        "46개의 토픽이 있으며, 각 토픽에 10개 이상의 example을 가지고 있음"
      ],
      "metadata": {
        "id": "tfLBCT4f_g8g"
      }
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Fy1Y7gzdaevF"
      },
      "source": [
        "**Loading the Reuters dataset**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 1,
      "metadata": {
        "id": "T24kkjvEaevF",
        "outputId": "0ba30662-4682-459f-8db1-d87f173c3a5f",
        "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",
            "2113536/2110848 [==============================] - 0s 0us/step\n",
            "2121728/2110848 [==============================] - 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": 2,
      "metadata": {
        "id": "nqmjxa6haevF",
        "outputId": "70126fad-d16d-48a5-d70f-3563559e8b3a",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "8982"
            ]
          },
          "metadata": {},
          "execution_count": 2
        }
      ],
      "source": [
        "len(train_data)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 3,
      "metadata": {
        "id": "0EE3x_NkaevG",
        "outputId": "d5a11f97-aa27-4940-c7dc-ec08ff3dad3e",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "2246"
            ]
          },
          "metadata": {},
          "execution_count": 3
        }
      ],
      "source": [
        "len(test_data)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 4,
      "metadata": {
        "id": "ftqQT2_EaevG",
        "outputId": "17f5aa8c-a59b-4439-e633-6bfd13322703",
        "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": 4
        }
      ],
      "source": [
        "train_data[10]"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "KtECD8zxaevG"
      },
      "source": [
        "**Decoding newswires back to text**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 5,
      "metadata": {
        "id": "P6teVZQaaevH",
        "outputId": "aa099f39-e6c8-4d55-a2b8-9c29fa53bdfe",
        "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",
            "557056/550378 [==============================] - 0s 0us/step\n",
            "565248/550378 [==============================] - 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": 6,
      "metadata": {
        "id": "WCu3t_4HaevH",
        "outputId": "d02705c9-97a4-49d5-ec13-d4d2192e6153",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "3"
            ]
          },
          "metadata": {},
          "execution_count": 6
        }
      ],
      "source": [
        "train_labels[10]"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "decoded_newswire"
      ],
      "metadata": {
        "id": "BtZ36Rz3_9p2",
        "outputId": "c825a236-b0da-4c54-d861-653065e595f1",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 89
        }
      },
      "execution_count": 9,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "'? ? ? said as a result of its december acquisition of space co it expects earnings per share in 1987 of 1 15 to 1 30 dlrs per share up from 70 cts in 1986 the company said pretax net should rise to nine to 10 mln dlrs from six mln dlrs in 1986 and rental operation revenues to 19 to 22 mln dlrs from 12 5 mln dlrs it said cash flow per share this year should be 2 50 to three dlrs reuter 3'"
            ],
            "application/vnd.google.colaboratory.intrinsic+json": {
              "type": "string"
            }
          },
          "metadata": {},
          "execution_count": 9
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "VNBJ5V9YaevI"
      },
      "source": [
        "### Preparing the data"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "7pcviomNaevI"
      },
      "source": [
        "**Encoding the input data**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 12,
      "metadata": {
        "id": "ifCfS8oPaevJ"
      },
      "outputs": [],
      "source": [
        "x_train = vectorize_sequences(train_data)\n",
        "x_test = vectorize_sequences(test_data)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "z5c3iYxlaevJ"
      },
      "source": [
        "**Encoding the labels**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 14,
      "metadata": {
        "id": "WNMLIzSqaevJ"
      },
      "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": 15,
      "metadata": {
        "id": "E3Sy5InAaevK"
      },
      "outputs": [],
      "source": [
        "# one hot encoding (keras 내장)\n",
        "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": "tKGwrOiJaevK"
      },
      "source": [
        "### Building your model"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "UDyOrQGyaevL"
      },
      "source": [
        "**Model definition**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 24,
      "metadata": {
        "id": "OOaJPe-qaevL"
      },
      "outputs": [],
      "source": [
        "model = keras.Sequential([\n",
        "    layers.Dense(64, activation=\"relu\"), # 조금 더 큰 층을 사용하자! 16차원 공간으로는 46차원 클래스를 구분하기에 너무 작음 \n",
        "    layers.Dense(64, activation=\"relu\"),\n",
        "    layers.Dense(46, activation=\"softmax\") # softmax함수의 output[i]은 i에 속할 prob며, 46개의 값을 모두 더하면 1이 된다 \n",
        "])"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "v9PQ0aoRaevM"
      },
      "source": [
        "**Compiling the model**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 20,
      "metadata": {
        "id": "dRe1i8QKaevM"
      },
      "outputs": [],
      "source": [
        "model.compile(optimizer=\"rmsprop\",\n",
        "              loss=\"categorical_crossentropy\", #  두 확률 분포 사이의 거리를 측정하는 loss function.\n",
        "              metrics=[\"accuracy\"])"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "PPRGj7V4aevN"
      },
      "source": [
        "### Validating your approach"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "PrdVth9VaevN"
      },
      "source": [
        "**Setting aside a validation set**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 21,
      "metadata": {
        "id": "1T_18ZDdaevN"
      },
      "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": "Int1Ig8uaevO"
      },
      "source": [
        "**Training the model**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 22,
      "metadata": {
        "id": "KavY3bWmaevO",
        "outputId": "012e274a-c642-4ea2-a521-2dc405caf6ed",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/20\n",
            "16/16 [==============================] - 2s 64ms/step - loss: 2.7331 - accuracy: 0.5247 - val_loss: 1.8661 - val_accuracy: 0.6360\n",
            "Epoch 2/20\n",
            "16/16 [==============================] - 1s 52ms/step - loss: 1.4837 - accuracy: 0.7035 - val_loss: 1.3390 - val_accuracy: 0.7160\n",
            "Epoch 3/20\n",
            "16/16 [==============================] - 1s 50ms/step - loss: 1.0751 - accuracy: 0.7795 - val_loss: 1.1495 - val_accuracy: 0.7550\n",
            "Epoch 4/20\n",
            "16/16 [==============================] - 1s 51ms/step - loss: 0.8368 - accuracy: 0.8281 - val_loss: 1.0556 - val_accuracy: 0.7710\n",
            "Epoch 5/20\n",
            "16/16 [==============================] - 1s 49ms/step - loss: 0.6637 - accuracy: 0.8624 - val_loss: 0.9694 - val_accuracy: 0.7950\n",
            "Epoch 6/20\n",
            "16/16 [==============================] - 1s 49ms/step - loss: 0.5314 - accuracy: 0.8908 - val_loss: 0.9372 - val_accuracy: 0.8010\n",
            "Epoch 7/20\n",
            "16/16 [==============================] - 1s 49ms/step - loss: 0.4217 - accuracy: 0.9131 - val_loss: 0.8984 - val_accuracy: 0.8110\n",
            "Epoch 8/20\n",
            "16/16 [==============================] - 1s 50ms/step - loss: 0.3456 - accuracy: 0.9266 - val_loss: 0.8887 - val_accuracy: 0.8120\n",
            "Epoch 9/20\n",
            "16/16 [==============================] - 1s 51ms/step - loss: 0.2819 - accuracy: 0.9390 - val_loss: 0.9268 - val_accuracy: 0.8140\n",
            "Epoch 10/20\n",
            "16/16 [==============================] - 1s 49ms/step - loss: 0.2436 - accuracy: 0.9440 - val_loss: 0.8995 - val_accuracy: 0.8210\n",
            "Epoch 11/20\n",
            "16/16 [==============================] - 1s 50ms/step - loss: 0.2064 - accuracy: 0.9501 - val_loss: 0.9139 - val_accuracy: 0.8150\n",
            "Epoch 12/20\n",
            "16/16 [==============================] - 1s 51ms/step - loss: 0.1846 - accuracy: 0.9519 - val_loss: 0.9988 - val_accuracy: 0.7980\n",
            "Epoch 13/20\n",
            "16/16 [==============================] - 1s 50ms/step - loss: 0.1643 - accuracy: 0.9523 - val_loss: 0.9787 - val_accuracy: 0.8130\n",
            "Epoch 14/20\n",
            "16/16 [==============================] - 1s 50ms/step - loss: 0.1527 - accuracy: 0.9544 - val_loss: 1.0050 - val_accuracy: 0.8090\n",
            "Epoch 15/20\n",
            "16/16 [==============================] - 1s 50ms/step - loss: 0.1370 - accuracy: 0.9560 - val_loss: 0.9859 - val_accuracy: 0.8100\n",
            "Epoch 16/20\n",
            "16/16 [==============================] - 1s 50ms/step - loss: 0.1341 - accuracy: 0.9563 - val_loss: 1.0301 - val_accuracy: 0.8030\n",
            "Epoch 17/20\n",
            "16/16 [==============================] - 1s 50ms/step - loss: 0.1257 - accuracy: 0.9553 - val_loss: 1.0881 - val_accuracy: 0.7940\n",
            "Epoch 18/20\n",
            "16/16 [==============================] - 1s 49ms/step - loss: 0.1177 - accuracy: 0.9578 - val_loss: 1.0784 - val_accuracy: 0.8060\n",
            "Epoch 19/20\n",
            "16/16 [==============================] - 1s 50ms/step - loss: 0.1148 - accuracy: 0.9580 - val_loss: 1.0546 - val_accuracy: 0.8070\n",
            "Epoch 20/20\n",
            "16/16 [==============================] - 1s 49ms/step - loss: 0.1134 - accuracy: 0.9549 - val_loss: 1.1865 - val_accuracy: 0.7880\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": "Q86bkksOaevO"
      },
      "source": [
        "**Plotting the training and validation loss**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 25,
      "metadata": {
        "id": "1tv464XmaevP",
        "outputId": "28d34949-7641-486c-f86d-76550ea13056",
        "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": [
        "import matplotlib.pyplot as plt\n",
        "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": "2s-GiZs9aevP"
      },
      "source": [
        "**Plotting the training and validation accuracy**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 26,
      "metadata": {
        "id": "z0jRPTtEaevQ",
        "outputId": "de216982-3a6f-4b12-ebef-6ca8f8c60cc2",
        "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",
      "source": [
        "epochs = 9 정도에서 오버피팅이 발생한다."
      ],
      "metadata": {
        "id": "qlzOgkjMEUGC"
      }
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "A7pnL8zUaevQ"
      },
      "source": [
        "**Retraining a model from scratch**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 27,
      "metadata": {
        "id": "DTPQdvbQaevR",
        "outputId": "72b54dcc-079a-47e5-a01c-1877ed532889",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/9\n",
            "18/18 [==============================] - 1s 44ms/step - loss: 2.4905 - accuracy: 0.4774\n",
            "Epoch 2/9\n",
            "18/18 [==============================] - 1s 43ms/step - loss: 1.3528 - accuracy: 0.7145\n",
            "Epoch 3/9\n",
            "18/18 [==============================] - 1s 44ms/step - loss: 1.0050 - accuracy: 0.7841\n",
            "Epoch 4/9\n",
            "18/18 [==============================] - 1s 44ms/step - loss: 0.7854 - accuracy: 0.8292\n",
            "Epoch 5/9\n",
            "18/18 [==============================] - 1s 43ms/step - loss: 0.6238 - accuracy: 0.8674\n",
            "Epoch 6/9\n",
            "18/18 [==============================] - 1s 43ms/step - loss: 0.4960 - accuracy: 0.8952\n",
            "Epoch 7/9\n",
            "18/18 [==============================] - 1s 44ms/step - loss: 0.3972 - accuracy: 0.9162\n",
            "Epoch 8/9\n",
            "18/18 [==============================] - 1s 43ms/step - loss: 0.3249 - accuracy: 0.9292\n",
            "Epoch 9/9\n",
            "18/18 [==============================] - 1s 44ms/step - loss: 0.2695 - accuracy: 0.9399\n",
            "71/71 [==============================] - 0s 3ms/step - loss: 0.9738 - accuracy: 0.7850\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": 28,
      "metadata": {
        "id": "bkv5v6udaevR",
        "outputId": "2f654572-751b-4bf8-84d4-8adcd5fb29f9",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[0.973783016204834, 0.7849510312080383]"
            ]
          },
          "metadata": {},
          "execution_count": 28
        }
      ],
      "source": [
        "results # 약 78%의 정확도"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 29,
      "metadata": {
        "id": "hoe-2Ok0aevS",
        "outputId": "794b3005-500e-489d-8ec5-35a926ef3b5c",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "0.18833481745325023"
            ]
          },
          "metadata": {},
          "execution_count": 29
        }
      ],
      "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()\n",
        "# 이건 뭘까....................................................................???"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "qkPcPknjaevS"
      },
      "source": [
        "### Generating predictions on new data"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 30,
      "metadata": {
        "id": "4jApcXzdaevT"
      },
      "outputs": [],
      "source": [
        "predictions = model.predict(x_test)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 31,
      "metadata": {
        "id": "BgoctvblaevT",
        "outputId": "f8c3d55c-6928-426c-983e-23a0ddb14270",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(46,)"
            ]
          },
          "metadata": {},
          "execution_count": 31
        }
      ],
      "source": [
        "predictions[0].shape"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "predictions[0]"
      ],
      "metadata": {
        "id": "1-F9M0SNFm4v",
        "outputId": "09f852f4-84c8-4782-c0e7-4c8ad8783475",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "execution_count": 38,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([6.78879906e-06, 5.03536321e-05, 2.03105174e-06, 9.16004002e-01,\n",
              "       8.01233575e-02, 1.99025635e-06, 7.78363483e-06, 3.07450573e-05,\n",
              "       7.22226163e-04, 4.94765686e-07, 5.19944042e-05, 2.58993299e-04,\n",
              "       7.64691067e-05, 8.53158417e-05, 2.22782219e-05, 2.37533859e-05,\n",
              "       5.29270736e-04, 1.20836550e-04, 2.56715848e-05, 5.20560483e-04,\n",
              "       2.72248610e-04, 3.77403107e-04, 1.01483856e-05, 1.10872425e-04,\n",
              "       4.36205710e-06, 2.72916295e-05, 2.81941129e-06, 3.40042316e-05,\n",
              "       2.68721506e-05, 8.38007982e-05, 2.07360208e-05, 3.40782462e-05,\n",
              "       2.03549334e-05, 4.08385768e-06, 1.96981655e-05, 9.27697693e-06,\n",
              "       1.43052948e-05, 1.79983181e-05, 6.07085076e-06, 6.57340424e-05,\n",
              "       8.55945109e-06, 1.47025188e-04, 7.34358935e-07, 1.41089731e-05,\n",
              "       2.20628863e-06, 2.75952488e-07], dtype=float32)"
            ]
          },
          "metadata": {},
          "execution_count": 38
        }
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 32,
      "metadata": {
        "id": "UOYF6LsUaevT",
        "outputId": "0acdc72f-2578-4541-ec6d-b893abbe2ded",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "1.0"
            ]
          },
          "metadata": {},
          "execution_count": 32
        }
      ],
      "source": [
        "np.sum(predictions[0])"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 39,
      "metadata": {
        "id": "36aqXlgqaevU",
        "outputId": "66ebf3d8-3f92-45b4-e925-62c4368ae999",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "3"
            ]
          },
          "metadata": {},
          "execution_count": 39
        }
      ],
      "source": [
        "np.argmax(predictions[0])"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "3gFUtIn7aevU"
      },
      "source": [
        "### A different way to handle the labels and the loss"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 40,
      "metadata": {
        "id": "bxUQXsokaevU"
      },
      "outputs": [],
      "source": [
        "y_train = np.array(train_labels)\n",
        "y_test = np.array(test_labels)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 41,
      "metadata": {
        "id": "6lCfml4kaevV"
      },
      "outputs": [],
      "source": [
        "model.compile(optimizer=\"rmsprop\",\n",
        "              loss=\"sparse_categorical_crossentropy\",\n",
        "              metrics=[\"accuracy\"])"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "BvF8FJI4aevV"
      },
      "source": [
        "### The importance of having sufficiently large intermediate layers"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "LoZlxyYWaevV"
      },
      "source": [
        "**A model with an information bottleneck**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 42,
      "metadata": {
        "id": "zrs_kSGWaevW",
        "outputId": "4f659b2a-5963-4baf-9d81-f6742f52107a",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/20\n",
            "63/63 [==============================] - 2s 21ms/step - loss: 3.0788 - accuracy: 0.0873 - val_loss: 2.3693 - val_accuracy: 0.3000\n",
            "Epoch 2/20\n",
            "63/63 [==============================] - 1s 18ms/step - loss: 1.8819 - accuracy: 0.5173 - val_loss: 1.6085 - val_accuracy: 0.6550\n",
            "Epoch 3/20\n",
            "63/63 [==============================] - 1s 18ms/step - loss: 1.3610 - accuracy: 0.6740 - val_loss: 1.4366 - val_accuracy: 0.6550\n",
            "Epoch 4/20\n",
            "63/63 [==============================] - 1s 18ms/step - loss: 1.1712 - accuracy: 0.6963 - val_loss: 1.3801 - val_accuracy: 0.6710\n",
            "Epoch 5/20\n",
            "63/63 [==============================] - 1s 18ms/step - loss: 1.0438 - accuracy: 0.7296 - val_loss: 1.3421 - val_accuracy: 0.6820\n",
            "Epoch 6/20\n",
            "63/63 [==============================] - 1s 18ms/step - loss: 0.9401 - accuracy: 0.7750 - val_loss: 1.3363 - val_accuracy: 0.6950\n",
            "Epoch 7/20\n",
            "63/63 [==============================] - 1s 22ms/step - loss: 0.8520 - accuracy: 0.7934 - val_loss: 1.3570 - val_accuracy: 0.7000\n",
            "Epoch 8/20\n",
            "63/63 [==============================] - 2s 24ms/step - loss: 0.7801 - accuracy: 0.8072 - val_loss: 1.3884 - val_accuracy: 0.7030\n",
            "Epoch 9/20\n",
            "63/63 [==============================] - 1s 19ms/step - loss: 0.7159 - accuracy: 0.8202 - val_loss: 1.4222 - val_accuracy: 0.7100\n",
            "Epoch 10/20\n",
            "63/63 [==============================] - 1s 19ms/step - loss: 0.6642 - accuracy: 0.8326 - val_loss: 1.4750 - val_accuracy: 0.7070\n",
            "Epoch 11/20\n",
            "63/63 [==============================] - 1s 19ms/step - loss: 0.6180 - accuracy: 0.8433 - val_loss: 1.5045 - val_accuracy: 0.7130\n",
            "Epoch 12/20\n",
            "63/63 [==============================] - 1s 19ms/step - loss: 0.5773 - accuracy: 0.8520 - val_loss: 1.5595 - val_accuracy: 0.7090\n",
            "Epoch 13/20\n",
            "63/63 [==============================] - 1s 18ms/step - loss: 0.5407 - accuracy: 0.8629 - val_loss: 1.6405 - val_accuracy: 0.7040\n",
            "Epoch 14/20\n",
            "63/63 [==============================] - 1s 19ms/step - loss: 0.5131 - accuracy: 0.8725 - val_loss: 1.6322 - val_accuracy: 0.7050\n",
            "Epoch 15/20\n",
            "63/63 [==============================] - 1s 18ms/step - loss: 0.4856 - accuracy: 0.8761 - val_loss: 1.7332 - val_accuracy: 0.7000\n",
            "Epoch 16/20\n",
            "63/63 [==============================] - 1s 18ms/step - loss: 0.4629 - accuracy: 0.8797 - val_loss: 1.7321 - val_accuracy: 0.7040\n",
            "Epoch 17/20\n",
            "63/63 [==============================] - 1s 19ms/step - loss: 0.4417 - accuracy: 0.8840 - val_loss: 1.8148 - val_accuracy: 0.6980\n",
            "Epoch 18/20\n",
            "63/63 [==============================] - 1s 18ms/step - loss: 0.4261 - accuracy: 0.8856 - val_loss: 1.8829 - val_accuracy: 0.6970\n",
            "Epoch 19/20\n",
            "63/63 [==============================] - 1s 19ms/step - loss: 0.4063 - accuracy: 0.8886 - val_loss: 1.9276 - val_accuracy: 0.6990\n",
            "Epoch 20/20\n",
            "63/63 [==============================] - 1s 18ms/step - loss: 0.3929 - accuracy: 0.8899 - val_loss: 1.9673 - val_accuracy: 0.6980\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<keras.callbacks.History at 0x7f86098e95d0>"
            ]
          },
          "metadata": {},
          "execution_count": 42
        }
      ],
      "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",
        "# categorical crossentropy는 레이블이 범주형 인코딩일 때 -> 정수 레이블을 사용할 때는 sparse_categorical_crossentropy를 사용\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": "xEqzFV_uaevW"
      },
      "source": [
        "### Further experiments"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "- 최종 분류 전 2개의 layor 사용 -> 1개 또는 3개의 layor 사용해보기\n",
        "- 더 많거나 적은 units 사용해보기"
      ],
      "metadata": {
        "id": "u9dvoe6GG2dN"
      }
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "fl4rHLWyaevW"
      },
      "source": [
        "### Wrapping up"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "80fwlF7QaevX"
      },
      "source": [
        "## Predicting house prices: A regression example"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "* regression : 예측 <-> logistic regression ; 분류"
      ],
      "metadata": {
        "id": "sSDsQ1JD_HEC"
      }
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "kRRTm9paaevX"
      },
      "source": [
        "### The Boston Housing Price dataset"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "PhbsZv1vaevX"
      },
      "source": [
        "**Loading the Boston housing dataset**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 43,
      "metadata": {
        "id": "AP20fcziaevY",
        "outputId": "1030b551-c179-4dfa-bde8-eac4cb9823b7",
        "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",
            "57344/57026 [==============================] - 0s 0us/step\n",
            "65536/57026 [==================================] - 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": 55,
      "metadata": {
        "id": "shly6xv9aevY",
        "outputId": "60f12e61-1f95-4f0f-bd22-42dc0b075632",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(404, 13)"
            ]
          },
          "metadata": {},
          "execution_count": 55
        }
      ],
      "source": [
        "train_data.shape"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 45,
      "metadata": {
        "id": "hUlte9bhaevY",
        "outputId": "e73895ab-b826-46d8-af8b-ea7f1d6776d8",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(102, 13)"
            ]
          },
          "metadata": {},
          "execution_count": 45
        }
      ],
      "source": [
        "test_data.shape"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 54,
      "metadata": {
        "id": "mBltVM7AaevZ",
        "outputId": "240c42ab-c99d-4f97-e702-fe8730658f77",
        "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": 54
        }
      ],
      "source": [
        "train_targets # 주택의 중간 가격, 1000$ 단위"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "jLUi2UXfaevZ"
      },
      "source": [
        "### Preparing the data"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "yJ8l63m7aevZ"
      },
      "source": [
        "**Normalizing the data**"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "different scale -> nomalizing 필요"
      ],
      "metadata": {
        "id": "AFv412UpI5Bt"
      }
    },
    {
      "cell_type": "code",
      "execution_count": 56,
      "metadata": {
        "id": "KSG1GyzSaeva"
      },
      "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\n",
        "# 반드시 train data의 mean과 std를 사용해 정규화 "
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "kTGYxKSYaeva"
      },
      "source": [
        "### Building your model"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "1J92-FRPaeva"
      },
      "source": [
        "**Model definition**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 59,
      "metadata": {
        "id": "qtgh9q_Eaevb"
      },
      "outputs": [],
      "source": [
        "def build_model():\n",
        "    model = keras.Sequential([\n",
        "        layers.Dense(64, activation=\"relu\"), # small trainign data -> overfitting 확률 높음 -> 작은 model\n",
        "        layers.Dense(64, activation=\"relu\"),\n",
        "        layers.Dense(1) # linear layor, scalar unit을 위한 구성 \n",
        "    ])\n",
        "    model.compile(optimizer=\"rmsprop\", loss=\"mse\", metrics=[\"mae\"]) # loss = mse : 회귀 문제에서 널리 사용됨 / matrics = mean absolute error : 예측과 타깃 사이의 절댓값\n",
        "    return model"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "SxFSTUmuaevb"
      },
      "source": [
        "### Validating your approach using K-fold validation"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "J1NiM3ODaevc"
      },
      "source": [
        "**K-fold validation**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 60,
      "metadata": {
        "id": "FH4MJ_eNaevc",
        "outputId": "c9239d3d-a8dd-4f45-e72d-10107de0258e",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Processing fold #0\n",
            "Processing fold #1\n",
            "Processing fold #2\n",
            "Processing fold #3\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\"Processing fold #{i}\")\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": 61,
      "metadata": {
        "id": "t0jcvj7Vaevc",
        "outputId": "f5f3b24f-1690-4024-9ac7-891045859893",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[2.2002570629119873, 2.4157137870788574, 2.393690586090088, 2.42057204246521]"
            ]
          },
          "metadata": {},
          "execution_count": 61
        }
      ],
      "source": [
        "all_scores"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 62,
      "metadata": {
        "id": "JuEtnl4Raevd",
        "outputId": "909013a5-bcb9-48fa-a948-d1e1eef7bdbc",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "2.3575583696365356"
            ]
          },
          "metadata": {},
          "execution_count": 62
        }
      ],
      "source": [
        "np.mean(all_scores)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "jPXb4xdRaevd"
      },
      "source": [
        "**Saving the validation logs at each fold**"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "ephochs을 500으로 늘려 훈련"
      ],
      "metadata": {
        "id": "-Drd2hUXLE-e"
      }
    },
    {
      "cell_type": "code",
      "execution_count": 63,
      "metadata": {
        "id": "kXGuS_Maaevd",
        "outputId": "daa4dc1f-9be0-4016-a354-119c4e2defa6",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Processing fold #0\n",
            "Processing fold #1\n",
            "Processing fold #2\n",
            "Processing fold #3\n"
          ]
        }
      ],
      "source": [
        "num_epochs = 500\n",
        "all_mae_histories = []\n",
        "for i in range(k):\n",
        "    print(f\"Processing fold #{i}\")\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": "hnssqvtCaeve"
      },
      "source": [
        "**Building the history of successive mean K-fold validation scores**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 64,
      "metadata": {
        "id": "pSPEE_o9aeve"
      },
      "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": "VgX9h_FUaeve"
      },
      "source": [
        "**Plotting validation scores**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 65,
      "metadata": {
        "id": "LGo2-O9waeve",
        "outputId": "a0e84975-8ef0-4afc-8cfc-40cafc16455e",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 281
        }
      },
      "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": "JsVKOiFraevf"
      },
      "source": [
        "**Plotting validation scores, excluding the first 10 data points**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 66,
      "metadata": {
        "id": "i3sULW-_aevf",
        "outputId": "396e12e4-2367-49ee-b4d7-35f87af4a057",
        "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": "lpgKrJl2aevf"
      },
      "source": [
        "**Training the final model**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 67,
      "metadata": {
        "id": "QfjK9if_aevg",
        "outputId": "aff5dc6b-eab2-4a95-bdde-5f11458ce518",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "4/4 [==============================] - 0s 3ms/step - loss: 14.9999 - mae: 2.5456\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": 68,
      "metadata": {
        "id": "D2McZ-AOaevg",
        "outputId": "153be2eb-1b8e-4aa7-8737-a31bc7e65333",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "2.545637607574463"
            ]
          },
          "metadata": {},
          "execution_count": 68
        }
      ],
      "source": [
        "test_mae_score"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "HqVXHFDMaevg"
      },
      "source": [
        "### Generating predictions on new data"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 69,
      "metadata": {
        "id": "cPuqc04laevh",
        "outputId": "c38b957f-b954-4fbb-dd15-0cec1705219d",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([8.024035], dtype=float32)"
            ]
          },
          "metadata": {},
          "execution_count": 69
        }
      ],
      "source": [
        "predictions = model.predict(test_data)\n",
        "predictions[0]"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "gy7BsIWhaevh"
      },
      "source": [
        "### Wrapping up"
      ]
    }
  ],
  "metadata": {
    "colab": {
      "collapsed_sections": [],
      "name": "chapter04_getting-started-with-neural-networks.i",
      "provenance": [],
      "toc_visible": true
    },
    "kernelspec": {
      "display_name": "Python 3",
      "language": "python",
      "name": "python3"
    },
    "language_info": {
      "codemirror_mode": {
        "name": "ipython",
        "version": 3
      },
      "file_extension": ".py",
      "mimetype": "text/x-python",
      "name": "python",
      "nbconvert_exporter": "python",
      "pygments_lexer": "ipython3",
      "version": "3.7.0"
    }
  },
  "nbformat": 4,
  "nbformat_minor": 0
}