material/deep_learning/template_attribute_data_binary_classification.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "template_attribute_data_binary_classification.ipynb",
      "provenance": [],
      "collapsed_sections": []
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "giMgHzNY8861"
      },
      "source": [
        "# 속성 데이터 2진 분류 Template"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "qDu8wpKXq7kV"
      },
      "source": [
        "# 데이터 준비"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ROMsisKxw8MC"
      },
      "source": [
        "import numpy as np\n",
        "import pandas as pd\n",
        "import matplotlib.pyplot as plt"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "gwMz9NaYwx5e",
        "outputId": "4309f2b2-fcc0-4f60-9c12-68cb088e205d"
      },
      "source": [
        "!wget https://raw.githubusercontent.com/dhrim/2022_WISET/main/material/deep_learning/data/sonar.csv"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "--2021-09-24 14:37:19--  https://raw.githubusercontent.com/dhrim/2022_WISET/main/material/deep_learning/data/sonar.csv\n",
            "Resolving raw.githubusercontent.com (raw.githubusercontent.com)... 185.199.111.133, 185.199.110.133, 185.199.109.133, ...\n",
            "Connecting to raw.githubusercontent.com (raw.githubusercontent.com)|185.199.111.133|:443... connected.\n",
            "HTTP request sent, awaiting response... 200 OK\n",
            "Length: 87776 (86K) [text/plain]\n",
            "Saving to: ‘sonar.csv’\n",
            "\n",
            "\rsonar.csv             0%[                    ]       0  --.-KB/s               \rsonar.csv           100%[===================>]  85.72K  --.-KB/s    in 0.007s  \n",
            "\n",
            "2021-09-24 14:37:19 (12.1 MB/s) - ‘sonar.csv’ saved [87776/87776]\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "VckhCYiwxAZj"
      },
      "source": [
        "raw_data = pd.read_csv(\"sonar.csv\")\n",
        "data = raw_data.to_numpy()\n",
        "np.random.shuffle(data)\n",
        "\n",
        "x = data[:,:60]\n",
        "y = data[:,60]\n",
        "\n",
        "split_index = int(len(x)*0.8)\n",
        "train_x, test_x = x[:split_index], x[split_index:]\n",
        "train_y, test_y = y[:split_index], y[split_index:]"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "JskspNe2Z5F5",
        "outputId": "be93433f-2bc0-4d07-afa3-11a4afcb20b4"
      },
      "source": [
        "print(train_x[:1])\n",
        "print(train_y[:5])"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[[0.079  0.0707 0.0352 0.166  0.133  0.0226 0.0771 0.2678 0.5664 0.6609\n",
            "  0.5002 0.2583 0.165  0.4347 0.4515 0.4579 0.3366 0.4    0.5325 0.901\n",
            "  0.9939 0.3689 0.1012 0.0248 0.2318 0.3981 0.2259 0.5247 0.6898 0.8316\n",
            "  0.4326 0.3741 0.5756 0.8043 0.7963 0.7174 0.7056 0.8148 0.7601 0.6034\n",
            "  0.4554 0.4729 0.4478 0.3722 0.4693 0.3839 0.0768 0.1467 0.0777 0.0469\n",
            "  0.0193 0.0298 0.039  0.0294 0.0175 0.0249 0.0141 0.0073 0.0025 0.0101]]\n",
            "[1. 0. 1. 1. 1.]\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "XR8Ck1Er6HmY",
        "outputId": "1f0324c2-47c6-4a88-da17-d4552d11c60b"
      },
      "source": [
        "print(train_x.shape)\n",
        "print(train_y.shape)\n",
        "print(test_x.shape)\n",
        "print(test_y.shape)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "(165, 60)\n",
            "(165,)\n",
            "(42, 60)\n",
            "(42,)\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "PvaZhPiSq5Gc"
      },
      "source": [
        "# Template"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "S-wBUWBH_mTQ",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 897
        },
        "outputId": "cff1b968-bfdd-43b8-a35d-1b8c9fcd1b38"
      },
      "source": [
        "import numpy as np\n",
        "import pandas as pd\n",
        "import matplotlib.pyplot as plt\n",
        "\n",
        "import tensorflow as tf\n",
        "\n",
        "from tensorflow import keras\n",
        "from tensorflow.keras import optimizers\n",
        "from tensorflow.keras.layers import Dense\n",
        "from tensorflow.keras.metrics import AUC, Precision, Recall\n",
        "\n",
        "model = keras.Sequential()\n",
        "model.add(Dense(10, activation='relu', input_shape=(60,)))\n",
        "model.add(Dense(10, activation='relu'))\n",
        "model.add(Dense(1, activation=\"sigmoid\"))\n",
        "\n",
        "model.compile(optimizer=\"Adam\", loss=\"binary_crossentropy\", metrics=[\"acc\", \"AUC\", Precision(), Recall()])\n",
        "model.summary()\n",
        "\n",
        "history = model.fit(train_x, train_y, epochs=200, verbose=1, batch_size=128, validation_split=0.1)\n",
        "\n",
        "plt.plot(history.history['loss'])\n",
        "plt.plot(history.history['val_loss'])\n",
        "plt.show()\n",
        "\n",
        "\n",
        "\n",
        "loss, acc, auc, precision, recall = model.evaluate(test_x, test_y)\n",
        "print(\"loss=\", loss)\n",
        "print(\"acc=\", acc)\n",
        "print(\"auc=\", auc)\n",
        "print(\"precision=\", precision)\n",
        "print(\"recall=\", recall)\n",
        "\n",
        "\n",
        "\n",
        "\n",
        "y_ = model.predict(test_x)\n",
        "predicted = (y_>0.5).astype(np.int16).squeeze()\n",
        "\n",
        "plt.plot(test_y, \"o\")\n",
        "plt.plot(predicted, \".\")\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Model: \"sequential_3\"\n",
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "dense_9 (Dense)              (None, 10)                610       \n",
            "_________________________________________________________________\n",
            "dense_10 (Dense)             (None, 10)                110       \n",
            "_________________________________________________________________\n",
            "dense_11 (Dense)             (None, 1)                 11        \n",
            "=================================================================\n",
            "Total params: 731\n",
            "Trainable params: 731\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "2/2 [==============================] - 0s 7ms/step - loss: 0.3919 - acc: 0.8095 - auc: 0.9050 - precision_3: 0.8000 - recall_3: 0.8696\n",
            "loss= 0.3919418752193451\n",
            "acc= 0.8095238208770752\n",
            "auc= 0.9050343632698059\n",
            "precision= 0.800000011920929\n",
            "recall= 0.8695651888847351\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "M9r4Tc-fadQi"
      },
      "source": [
        ""
      ],
      "execution_count": null,
      "outputs": []
    }
  ]
}