material/deep_learning/treating_imbalanced_data.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "treating_imbalanced_data.ipynb",
      "provenance": [],
      "collapsed_sections": [],
      "toc_visible": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "code",
      "source": [
        "import numpy as np\n",
        "import pandas as pd\n",
        "import matplotlib.pyplot as plt\n",
        "import seaborn as sns\n",
        "from pandas import DataFrame\n",
        "\n",
        "import tensorflow as tf\n",
        "from tensorflow import keras\n",
        "\n",
        "from tensorflow.keras import optimizers\n",
        "from tensorflow.keras.layers import Dense, Input\n",
        "from tensorflow.keras import Sequential"
      ],
      "metadata": {
        "id": "BXZSqC0Od0bY"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "# 데이터 준비"
      ],
      "metadata": {
        "id": "JBDsVPeXjutv"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "from sklearn.datasets import make_moons\n",
        "\n",
        "\n",
        "# 데이터 생성\n",
        "base_x, base_y = make_moons(n_samples=(1000,50), noise=0.2, random_state=32)\n",
        "\n",
        "\n",
        "# 데이터 그림 그리기\n",
        "df = DataFrame(dict(x=base_x[:,0], y=base_x[:,1], label=base_y))\n",
        "colors = {0:'red', 1:'blue'}\n",
        "fig, ax = plt.subplots()\n",
        "grouped = df.groupby('label')\n",
        "for key, group in grouped:\n",
        "    group.plot(ax=ax, kind='scatter', x='x', y='y', label=key, color=colors[key])\n",
        "plt.show()\n"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 279
        },
        "id": "BcFKzVpadvAR",
        "outputId": "116c4134-2fba-439a-856b-b877e793e1f5"
      },
      "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": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# 공통 실행 함수"
      ],
      "metadata": {
        "id": "psWnHrlxq3gV"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "from sklearn.model_selection import train_test_split\n",
        "from sklearn.metrics import classification_report\n",
        "from sklearn.metrics import confusion_matrix\n",
        "from sklearn.metrics import classification_report\n",
        "\n",
        "CLASS_0_LABEL = 0\n",
        "CLASS_1_LABEL = 1\n",
        "DEFAULT_CLASS_WEIGHT = {CLASS_0_LABEL:1.0, CLASS_1_LABEL:1.0}\n",
        "\n",
        "def do_classify(x, y, class_weight=DEFAULT_CLASS_WEIGHT):\n",
        "\n",
        "    train_x, test_x, train_y, test_y = train_test_split(x, y, test_size=0.3)\n",
        "\n",
        "    model = Sequential()\n",
        "    model.add(Input(2))\n",
        "    model.add(Dense(20, activation='relu'))\n",
        "    model.add(Dense(20, activation='relu'))\n",
        "    model.add(Dense(2, activation=\"softmax\"))\n",
        "    model.compile(optimizer='Adam', loss='sparse_categorical_crossentropy', metrics=[\"acc\"])\n",
        "\n",
        "    base_history = model.fit(\n",
        "        train_x, train_y, epochs=100, batch_size=128, verbose=0, validation_split=0.2,\n",
        "        class_weight=class_weight\n",
        "    )\n",
        "\n",
        "\n",
        "    plt.plot(base_history.history['loss'], label='train_loss')\n",
        "    plt.plot(base_history.history['val_loss'], label='val_loss')\n",
        "    plt.legend()\n",
        "    plt.show()\n",
        "\n",
        "    loss, acc = model.evaluate(test_x, test_y)\n",
        "    print(\"loss=\", loss)\n",
        "    print(\"acc=\", acc)\n",
        "\n",
        "\n",
        "    y_ = model.predict(test_x)\n",
        "    predicted = np.argmax(y_, axis=-1)\n",
        "\n",
        "\n",
        "    cm = confusion_matrix(test_y, predicted)\n",
        "    print(cm)\n",
        "    sns.heatmap(cm, annot=True, fmt=\"d\")\n",
        "\n",
        "    print(classification_report(test_y, predicted))\n",
        "    "
      ],
      "metadata": {
        "id": "y2_lg1lsmvyh"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "# 처리 하지 않았을 때"
      ],
      "metadata": {
        "id": "_js1lcQFjsoB"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "x = base_x.copy()\n",
        "y = base_y.copy()\n",
        "\n",
        "do_classify(x, y)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 760
        },
        "id": "yy7sk-GEms2R",
        "outputId": "35d62854-b2a6-4a8c-b940-51d96546372c"
      },
      "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": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "10/10 [==============================] - 0s 2ms/step - loss: 0.0837 - acc: 0.9683\n",
            "loss= 0.08371222764253616\n",
            "acc= 0.9682539701461792\n",
            "[[300   4]\n",
            " [  6   5]]\n",
            "              precision    recall  f1-score   support\n",
            "\n",
            "           0       0.98      0.99      0.98       304\n",
            "           1       0.56      0.45      0.50        11\n",
            "\n",
            "    accuracy                           0.97       315\n",
            "   macro avg       0.77      0.72      0.74       315\n",
            "weighted avg       0.97      0.97      0.97       315\n",
            "\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 2 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# 오버샘플링"
      ],
      "metadata": {
        "id": "zvSo0M4CjxsU"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "x = base_x.copy()\n",
        "y = base_y.copy()"
      ],
      "metadata": {
        "id": "-3j5qyVYkfVX"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "class0_x = x[y==0]\n",
        "class0_y = y[y==0]\n",
        "class1_x = x[y==1]\n",
        "class1_y = y[y==1]\n",
        "\n",
        "print(len(class0_x), len(class1_x))\n",
        "imbalance_ratio = int(len(class0_x)/len(class1_x))\n",
        "print(imbalance_ratio)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ZeGsUTTAdutu",
        "outputId": "4a239f8b-e330-4f7f-c287-c45e1d7e2c99"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "1000 50\n",
            "20\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "class1_x = np.repeat(class1_x, imbalance_ratio, axis=0)\n",
        "class1_y = np.repeat(class1_y, imbalance_ratio, axis=0)\n",
        "print(len(class1_x))"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "2fJVv7x1durK",
        "outputId": "e53d13c0-5c05-4f08-b36f-127eda126137"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "1000\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "print(x.shape)\n",
        "print(class1_x.shape)\n",
        "print(y.shape)\n",
        "print(class1_y.shape)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "02OqivCmduk9",
        "outputId": "234a7e0c-09eb-4498-8b1e-1c4568f286dd"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "(1050, 2)\n",
            "(1000, 2)\n",
            "(1050,)\n",
            "(1000,)\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "x = np.vstack([class0_x, class1_x])\n",
        "y = np.hstack([class0_y, class1_y])\n",
        "print(x.shape)\n",
        "print(y.shape)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Bz64kcBil2_s",
        "outputId": "12132c59-24b9-4613-c143-647df46147c7"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "(2000, 2)\n",
            "(2000,)\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "do_classify(x, y)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 756
        },
        "id": "vebV5awImqRv",
        "outputId": "3d708654-038b-4480-cb2d-1bb4b7009553"
      },
      "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": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "19/19 [==============================] - 0s 2ms/step - loss: 0.1778 - acc: 0.9367\n",
            "loss= 0.17778272926807404\n",
            "acc= 0.9366666674613953\n",
            "[[280  14]\n",
            " [ 24 282]]\n",
            "              precision    recall  f1-score   support\n",
            "\n",
            "           0       0.92      0.95      0.94       294\n",
            "           1       0.95      0.92      0.94       306\n",
            "\n",
            "    accuracy                           0.94       600\n",
            "   macro avg       0.94      0.94      0.94       600\n",
            "weighted avg       0.94      0.94      0.94       600\n",
            "\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 2 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        ""
      ],
      "metadata": {
        "id": "hKerD0sYmqO9"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "# 언더샘플링"
      ],
      "metadata": {
        "id": "z_gObmFtne9z"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "x = base_x.copy()\n",
        "y = base_y.copy()"
      ],
      "metadata": {
        "id": "7RyEjX1une96"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "class0_x = x[y==0]\n",
        "class0_y = y[y==0]\n",
        "class1_x = x[y==1]\n",
        "class1_y = y[y==1]\n",
        "\n",
        "print(len(class0_x), len(class1_x))\n",
        "class1_count = len(class1_x)\n",
        "print(class1_count)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "aa0739b7-ea83-4767-dffb-c4850e851ffe",
        "id": "gpQHxVFUne97"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "1000 50\n",
            "50\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "class0_x = class0_x[:class1_count]\n",
        "class0_y = class0_y[:class1_count]\n",
        "\n",
        "print(len(class0_x), len(class1_x))"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "b1db8b6a-c45f-450a-e20e-144a331d4f5a",
        "id": "M3DrIAeDne97"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "50 50\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "x = np.vstack([class0_x, class1_x])\n",
        "y = np.hstack([class0_y, class1_y])\n",
        "print(x.shape)\n",
        "print(y.shape)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "b83c7586-1804-4d6b-f695-56ff211cfcca",
        "id": "oRrLGTzyne98"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "(100, 2)\n",
            "(100,)\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "do_classify(x, y)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 760
        },
        "outputId": "1c486c6d-69cd-4d8f-fa94-27755c5fed2a",
        "id": "9NLJHIunne98"
      },
      "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": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "1/1 [==============================] - 0s 18ms/step - loss: 0.5036 - acc: 0.7333\n",
            "loss= 0.5035920739173889\n",
            "acc= 0.7333333492279053\n",
            "[[10  3]\n",
            " [ 5 12]]\n",
            "              precision    recall  f1-score   support\n",
            "\n",
            "           0       0.67      0.77      0.71        13\n",
            "           1       0.80      0.71      0.75        17\n",
            "\n",
            "    accuracy                           0.73        30\n",
            "   macro avg       0.73      0.74      0.73        30\n",
            "weighted avg       0.74      0.73      0.73        30\n",
            "\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 2 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# class weight"
      ],
      "metadata": {
        "id": "a-YXDLeCobl4"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "x = base_x.copy()\n",
        "y = base_y.copy()"
      ],
      "metadata": {
        "id": "r7Y8jKlQpA5I"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "from sklearn.utils.class_weight import compute_class_weight\n",
        "classes = np.unique(y)\n",
        "print(classes)\n",
        "\n",
        "class_weight_ratio = compute_class_weight(\n",
        "    class_weight=\"balanced\", classes=classes, y=y.squeeze())\n",
        "print(class_weight_ratio)\n",
        "\n",
        "class_weight = { k:v for k, v in zip(classes, class_weight_ratio) }\n",
        "print(class_weight)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Fehe02nJobY8",
        "outputId": "a392052c-5e83-4cc0-ae2e-20e0708e6d9f"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[0 1]\n",
            "[ 0.525 10.5  ]\n",
            "{0: 0.525, 1: 10.5}\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "do_classify(x, y, class_weight)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 756
        },
        "id": "P4Dfa_00pqvY",
        "outputId": "241ef949-5443-4bcb-97dc-9ce11445e96d"
      },
      "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": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "10/10 [==============================] - 0s 2ms/step - loss: 0.3229 - acc: 0.8952\n",
            "loss= 0.32285207509994507\n",
            "acc= 0.8952381014823914\n",
            "[[275  31]\n",
            " [  2   7]]\n",
            "              precision    recall  f1-score   support\n",
            "\n",
            "           0       0.99      0.90      0.94       306\n",
            "           1       0.18      0.78      0.30         9\n",
            "\n",
            "    accuracy                           0.90       315\n",
            "   macro avg       0.59      0.84      0.62       315\n",
            "weighted avg       0.97      0.90      0.92       315\n",
            "\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 2 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    }
  ]
}