大數據分析期末報告3_正式版.ipynb

{
  "cells": [
    {
      "cell_type": "code",
      "execution_count": 8,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "3F_Zoz8z3YbA",
        "outputId": "90db7d84-0938-42d6-d9f5-19511ff4526f"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "+----+------------------+-------------------+----------------+------------------+-------------------+-------------------+-------------------+------------------+------------------+-------------------+------------------+------------------+------------------+------------------+------------------+------------------+------------------+-------------------+------------------+-------------------+--------------------+-------------------+------------------+------------------+------------------+------------------+--------------------+-------------------+------+-----+\n",
            "|Time|                V1|                 V2|              V3|                V4|                 V5|                 V6|                 V7|                V8|                V9|                V10|               V11|               V12|               V13|               V14|               V15|               V16|               V17|                V18|               V19|                V20|                 V21|                V22|               V23|               V24|               V25|               V26|                 V27|                V28|Amount|Class|\n",
            "+----+------------------+-------------------+----------------+------------------+-------------------+-------------------+-------------------+------------------+------------------+-------------------+------------------+------------------+------------------+------------------+------------------+------------------+------------------+-------------------+------------------+-------------------+--------------------+-------------------+------------------+------------------+------------------+------------------+--------------------+-------------------+------+-----+\n",
            "| 0.0|  -1.3598071336738|-0.0727811733098497|2.53634673796914|  1.37815522427443| -0.338320769942518|  0.462387777762292|  0.239598554061257|0.0986979012610507| 0.363786969611213| 0.0907941719789316|-0.551599533260813|-0.617800855762348|-0.991389847235408|-0.311169353699879|  1.46817697209427|-0.470400525259478| 0.207971241929242| 0.0257905801985591| 0.403992960255733|  0.251412098239705|  -0.018306777944153|  0.277837575558899|-0.110473910188767|0.0669280749146731| 0.128539358273528|-0.189114843888824|   0.133558376740387|-0.0210530534538215|149.62|    0|\n",
            "| 0.0|  1.19185711131486|   0.26615071205963|0.16648011335321| 0.448154078460911| 0.0600176492822243|-0.0823608088155687|-0.0788029833323113|0.0851016549148104|-0.255425128109186| -0.166974414004614|  1.61272666105479|  1.06523531137287|  0.48909501589608|-0.143772296441519| 0.635558093258208| 0.463917041022171|-0.114804663102346| -0.183361270123994|-0.145783041325259|-0.0690831352230203|  -0.225775248033138| -0.638671952771851| 0.101288021253234|-0.339846475529127| 0.167170404418143| 0.125894532368176|-0.00898309914322813| 0.0147241691924927|  2.69|    0|\n",
            "| 1.0| -1.35835406159823|  -1.34016307473609|1.77320934263119| 0.379779593034328| -0.503198133318193|   1.80049938079263|  0.791460956450422| 0.247675786588991| -1.51465432260583|  0.207642865216696| 0.624501459424895| 0.066083685268831| 0.717292731410831|-0.165945922763554|  2.34586494901581| -2.89008319444231|  1.10996937869599| -0.121359313195888| -2.26185709530414|  0.524979725224404|   0.247998153469754|  0.771679401917229| 0.909412262347719|-0.689280956490685|-0.327641833735251|-0.139096571514147| -0.0553527940384261|-0.0597518405929204|378.66|    0|\n",
            "| 1.0|-0.966271711572087| -0.185226008082898|1.79299333957872|-0.863291275036453|-0.0103088796030823|   1.24720316752486|   0.23760893977178| 0.377435874652262| -1.38702406270197|-0.0549519224713749|-0.226487263835401| 0.178228225877303| 0.507756869957169| -0.28792374549456|-0.631418117709045|  -1.0596472454325|-0.684092786345479|   1.96577500349538|  -1.2326219700892| -0.208037781160366|  -0.108300452035545|0.00527359678253453|-0.190320518742841| -1.17557533186321| 0.647376034602038|-0.221928844458407|  0.0627228487293033| 0.0614576285006353| 123.5|    0|\n",
            "| 2.0| -1.15823309349523|  0.877736754848451|  1.548717846511| 0.403033933955121| -0.407193377311653| 0.0959214624684256|  0.592940745385545|-0.270532677192282| 0.817739308235294|  0.753074431976354|-0.822842877946363|  0.53819555014995|   1.3458515932154| -1.11966983471731| 0.175121130008994|-0.451449182813529|-0.237033239362776|-0.0381947870352842| 0.803486924960175|  0.408542360392758|-0.00943069713232919|   0.79827849458971|-0.137458079619063| 0.141266983824769|-0.206009587619756| 0.502292224181569|   0.219422229513348|  0.215153147499206| 69.99|    0|\n",
            "+----+------------------+-------------------+----------------+------------------+-------------------+-------------------+-------------------+------------------+------------------+-------------------+------------------+------------------+------------------+------------------+------------------+------------------+------------------+-------------------+------------------+-------------------+--------------------+-------------------+------------------+------------------+------------------+------------------+--------------------+-------------------+------+-----+\n",
            "only showing top 5 rows\n",
            "\n",
            "root\n",
            " |-- Time: double (nullable = true)\n",
            " |-- V1: double (nullable = true)\n",
            " |-- V2: double (nullable = true)\n",
            " |-- V3: double (nullable = true)\n",
            " |-- V4: double (nullable = true)\n",
            " |-- V5: double (nullable = true)\n",
            " |-- V6: double (nullable = true)\n",
            " |-- V7: double (nullable = true)\n",
            " |-- V8: double (nullable = true)\n",
            " |-- V9: double (nullable = true)\n",
            " |-- V10: double (nullable = true)\n",
            " |-- V11: double (nullable = true)\n",
            " |-- V12: double (nullable = true)\n",
            " |-- V13: double (nullable = true)\n",
            " |-- V14: double (nullable = true)\n",
            " |-- V15: double (nullable = true)\n",
            " |-- V16: double (nullable = true)\n",
            " |-- V17: double (nullable = true)\n",
            " |-- V18: double (nullable = true)\n",
            " |-- V19: double (nullable = true)\n",
            " |-- V20: double (nullable = true)\n",
            " |-- V21: double (nullable = true)\n",
            " |-- V22: double (nullable = true)\n",
            " |-- V23: double (nullable = true)\n",
            " |-- V24: double (nullable = true)\n",
            " |-- V25: double (nullable = true)\n",
            " |-- V26: double (nullable = true)\n",
            " |-- V27: double (nullable = true)\n",
            " |-- V28: double (nullable = true)\n",
            " |-- Amount: double (nullable = true)\n",
            " |-- Class: integer (nullable = true)\n",
            "\n",
            "AUC-ROC 分數：0.9773\n",
            "+-----+----------+-----+\n",
            "|Class|prediction|count|\n",
            "+-----+----------+-----+\n",
            "|    1|       0.0|   40|\n",
            "|    0|       0.0|56867|\n",
            "|    1|       1.0|   58|\n",
            "|    0|       1.0|    8|\n",
            "+-----+----------+-----+\n",
            "\n"
          ]
        }
      ],
      "source": [
        "# 導入必要的庫\n",
        "from pyspark.sql import SparkSession\n",
        "from pyspark.sql.functions import col, when, explode, array\n",
        "from pyspark.ml.feature import VectorAssembler\n",
        "from pyspark.ml.classification import LinearSVC\n",
        "from pyspark.ml.evaluation import BinaryClassificationEvaluator\n",
        "from pyspark.ml.stat import Correlation\n",
        "\n",
        "# 初始化 SparkSession\n",
        "spark = SparkSession.builder.appName(\"CreditCardFraudDetection\").getOrCreate()\n",
        "\n",
        "# 讀取數據集\n",
        "file_path = \"/content/creditcard.csv\"  # 在 Colab 中設置文件路徑\n",
        "data = spark.read.csv(file_path, header=True, inferSchema=True)\n",
        "\n",
        "# 數據簡要查看\n",
        "data.show(5)\n",
        "data.printSchema()\n",
        "\n",
        "# 轉換標籤列（1 表示詐欺，0 表示正常交易）\n",
        "data = data.withColumn(\"Class\", when(col(\"Class\") == 1, 1).otherwise(0))\n",
        "\n",
        "# 特徵組合\n",
        "assembler = VectorAssembler(inputCols=[col for col in data.columns if col != \"Class\"], outputCol=\"features\")\n",
        "data = assembler.transform(data)\n",
        "\n",
        "# 訓練集和測試集劃分\n",
        "train_data, test_data = data.randomSplit([0.8, 0.2], seed=42)\n",
        "\n",
        "# 使用 Spark ML 的支持向量機\n",
        "svm = LinearSVC(featuresCol=\"features\", labelCol=\"Class\", maxIter=10, regParam=0.1)\n",
        "model = svm.fit(train_data)\n",
        "\n",
        "# 模型預測\n",
        "predictions = model.transform(test_data)\n",
        "\n",
        "# 評估模型性能\n",
        "evaluator = BinaryClassificationEvaluator(labelCol=\"Class\", rawPredictionCol=\"rawPrediction\", metricName=\"areaUnderROC\")\n",
        "auc = evaluator.evaluate(predictions)\n",
        "print(f\"AUC-ROC 分數：{auc:.4f}\")\n",
        "\n",
        "# 混淆矩陣計算\n",
        "predictions.groupBy(\"Class\", \"prediction\").count().show()\n",
        "\n",
        "# 結束 SparkSession\n",
        "spark.stop()\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 9,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "id": "u6aansO85EYr",
        "outputId": "631d2f89-79a7-4c30-f3cc-3935725caf73"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "+----+------------------+-------------------+----------------+------------------+-------------------+-------------------+-------------------+------------------+------------------+-------------------+------------------+------------------+------------------+------------------+------------------+------------------+------------------+-------------------+------------------+-------------------+--------------------+-------------------+------------------+------------------+------------------+------------------+--------------------+-------------------+------+-----+\n",
            "|Time|                V1|                 V2|              V3|                V4|                 V5|                 V6|                 V7|                V8|                V9|                V10|               V11|               V12|               V13|               V14|               V15|               V16|               V17|                V18|               V19|                V20|                 V21|                V22|               V23|               V24|               V25|               V26|                 V27|                V28|Amount|Class|\n",
            "+----+------------------+-------------------+----------------+------------------+-------------------+-------------------+-------------------+------------------+------------------+-------------------+------------------+------------------+------------------+------------------+------------------+------------------+------------------+-------------------+------------------+-------------------+--------------------+-------------------+------------------+------------------+------------------+------------------+--------------------+-------------------+------+-----+\n",
            "| 0.0|  -1.3598071336738|-0.0727811733098497|2.53634673796914|  1.37815522427443| -0.338320769942518|  0.462387777762292|  0.239598554061257|0.0986979012610507| 0.363786969611213| 0.0907941719789316|-0.551599533260813|-0.617800855762348|-0.991389847235408|-0.311169353699879|  1.46817697209427|-0.470400525259478| 0.207971241929242| 0.0257905801985591| 0.403992960255733|  0.251412098239705|  -0.018306777944153|  0.277837575558899|-0.110473910188767|0.0669280749146731| 0.128539358273528|-0.189114843888824|   0.133558376740387|-0.0210530534538215|149.62|    0|\n",
            "| 0.0|  1.19185711131486|   0.26615071205963|0.16648011335321| 0.448154078460911| 0.0600176492822243|-0.0823608088155687|-0.0788029833323113|0.0851016549148104|-0.255425128109186| -0.166974414004614|  1.61272666105479|  1.06523531137287|  0.48909501589608|-0.143772296441519| 0.635558093258208| 0.463917041022171|-0.114804663102346| -0.183361270123994|-0.145783041325259|-0.0690831352230203|  -0.225775248033138| -0.638671952771851| 0.101288021253234|-0.339846475529127| 0.167170404418143| 0.125894532368176|-0.00898309914322813| 0.0147241691924927|  2.69|    0|\n",
            "| 1.0| -1.35835406159823|  -1.34016307473609|1.77320934263119| 0.379779593034328| -0.503198133318193|   1.80049938079263|  0.791460956450422| 0.247675786588991| -1.51465432260583|  0.207642865216696| 0.624501459424895| 0.066083685268831| 0.717292731410831|-0.165945922763554|  2.34586494901581| -2.89008319444231|  1.10996937869599| -0.121359313195888| -2.26185709530414|  0.524979725224404|   0.247998153469754|  0.771679401917229| 0.909412262347719|-0.689280956490685|-0.327641833735251|-0.139096571514147| -0.0553527940384261|-0.0597518405929204|378.66|    0|\n",
            "| 1.0|-0.966271711572087| -0.185226008082898|1.79299333957872|-0.863291275036453|-0.0103088796030823|   1.24720316752486|   0.23760893977178| 0.377435874652262| -1.38702406270197|-0.0549519224713749|-0.226487263835401| 0.178228225877303| 0.507756869957169| -0.28792374549456|-0.631418117709045|  -1.0596472454325|-0.684092786345479|   1.96577500349538|  -1.2326219700892| -0.208037781160366|  -0.108300452035545|0.00527359678253453|-0.190320518742841| -1.17557533186321| 0.647376034602038|-0.221928844458407|  0.0627228487293033| 0.0614576285006353| 123.5|    0|\n",
            "| 2.0| -1.15823309349523|  0.877736754848451|  1.548717846511| 0.403033933955121| -0.407193377311653| 0.0959214624684256|  0.592940745385545|-0.270532677192282| 0.817739308235294|  0.753074431976354|-0.822842877946363|  0.53819555014995|   1.3458515932154| -1.11966983471731| 0.175121130008994|-0.451449182813529|-0.237033239362776|-0.0381947870352842| 0.803486924960175|  0.408542360392758|-0.00943069713232919|   0.79827849458971|-0.137458079619063| 0.141266983824769|-0.206009587619756| 0.502292224181569|   0.219422229513348|  0.215153147499206| 69.99|    0|\n",
            "+----+------------------+-------------------+----------------+------------------+-------------------+-------------------+-------------------+------------------+------------------+-------------------+------------------+------------------+------------------+------------------+------------------+------------------+------------------+-------------------+------------------+-------------------+--------------------+-------------------+------------------+------------------+------------------+------------------+--------------------+-------------------+------+-----+\n",
            "only showing top 5 rows\n",
            "\n",
            "root\n",
            " |-- Time: double (nullable = true)\n",
            " |-- V1: double (nullable = true)\n",
            " |-- V2: double (nullable = true)\n",
            " |-- V3: double (nullable = true)\n",
            " |-- V4: double (nullable = true)\n",
            " |-- V5: double (nullable = true)\n",
            " |-- V6: double (nullable = true)\n",
            " |-- V7: double (nullable = true)\n",
            " |-- V8: double (nullable = true)\n",
            " |-- V9: double (nullable = true)\n",
            " |-- V10: double (nullable = true)\n",
            " |-- V11: double (nullable = true)\n",
            " |-- V12: double (nullable = true)\n",
            " |-- V13: double (nullable = true)\n",
            " |-- V14: double (nullable = true)\n",
            " |-- V15: double (nullable = true)\n",
            " |-- V16: double (nullable = true)\n",
            " |-- V17: double (nullable = true)\n",
            " |-- V18: double (nullable = true)\n",
            " |-- V19: double (nullable = true)\n",
            " |-- V20: double (nullable = true)\n",
            " |-- V21: double (nullable = true)\n",
            " |-- V22: double (nullable = true)\n",
            " |-- V23: double (nullable = true)\n",
            " |-- V24: double (nullable = true)\n",
            " |-- V25: double (nullable = true)\n",
            " |-- V26: double (nullable = true)\n",
            " |-- V27: double (nullable = true)\n",
            " |-- V28: double (nullable = true)\n",
            " |-- Amount: double (nullable = true)\n",
            " |-- Class: integer (nullable = true)\n",
            "\n"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.10/dist-packages/xgboost/core.py:158: UserWarning: [00:57:55] WARNING: /workspace/src/learner.cc:740: \n",
            "Parameters: { \"use_label_encoder\" } are not used.\n",
            "\n",
            "  warnings.warn(smsg, UserWarning)\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "混淆矩陣：\n",
            "[[56857     7]\n",
            " [   19    79]]\n",
            "分類報告：\n",
            "              precision    recall  f1-score   support\n",
            "\n",
            "           0       1.00      1.00      1.00     56864\n",
            "           1       0.92      0.81      0.86        98\n",
            "\n",
            "    accuracy                           1.00     56962\n",
            "   macro avg       0.96      0.90      0.93     56962\n",
            "weighted avg       1.00      1.00      1.00     56962\n",
            "\n",
            "AUC-ROC 分數：0.9743\n"
          ]
        },
        {
          "data": {
            "image/png": "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",
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        }
      ],
      "source": [
        "# 導入必要的庫\n",
        "from pyspark.sql import SparkSession\n",
        "from pyspark.sql.functions import col, when\n",
        "from pyspark.ml.feature import VectorAssembler\n",
        "from pyspark.ml.evaluation import BinaryClassificationEvaluator\n",
        "\n",
        "import pandas as pd\n",
        "import matplotlib.pyplot as plt\n",
        "import seaborn as sns\n",
        "import xgboost as xgb\n",
        "\n",
        "# sklearn 模組\n",
        "from sklearn.model_selection import train_test_split\n",
        "from sklearn.metrics import classification_report, confusion_matrix, roc_auc_score\n",
        "\n",
        "# 初始化 SparkSession\n",
        "spark = SparkSession.builder.appName(\"CreditCardFraudDetection_XGBoost\").getOrCreate()\n",
        "\n",
        "# 讀取數據集\n",
        "file_path = \"/content/creditcard.csv\"  # 在 Colab 中設置文件路徑\n",
        "data = spark.read.csv(file_path, header=True, inferSchema=True)\n",
        "\n",
        "# 數據簡要查看\n",
        "data.show(5)\n",
        "data.printSchema()\n",
        "\n",
        "# 轉換標籤列（1 表示詐欺，0 表示正常交易）\n",
        "data = data.withColumn(\"Class\", when(col(\"Class\") == 1, 1).otherwise(0))\n",
        "\n",
        "# 特徵組合\n",
        "assembler = VectorAssembler(inputCols=[col for col in data.columns if col != \"Class\"], outputCol=\"features\")\n",
        "data = assembler.transform(data)\n",
        "\n",
        "# 將 Spark DataFrame 轉換為 Pandas DataFrame\n",
        "pandas_data = data.select([col for col in data.columns if col != \"features\"]).toPandas()\n",
        "X = pandas_data.drop(columns=[\"Class\"])\n",
        "y = pandas_data[\"Class\"]\n",
        "\n",
        "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42, stratify=y)\n",
        "\n",
        "# 使用 XGBoost 進行分類\n",
        "xgb_model = xgb.XGBClassifier(use_label_encoder=False, eval_metric='auc', random_state=42)\n",
        "xgb_model.fit(X_train, y_train)\n",
        "\n",
        "# 模型預測\n",
        "y_pred = xgb_model.predict(X_test)\n",
        "y_pred_prob = xgb_model.predict_proba(X_test)[:, 1]\n",
        "\n",
        "# 混淆矩陣\n",
        "conf_matrix = confusion_matrix(y_test, y_pred)\n",
        "print(\"混淆矩陣：\")\n",
        "print(conf_matrix)\n",
        "\n",
        "# 混淆矩陣可視化\n",
        "plt.figure(figsize=(8, 6))\n",
        "sns.heatmap(conf_matrix, annot=True, fmt=\"d\", cmap=\"Blues\", xticklabels=[\"Non-Fraud\", \"Fraud\"], yticklabels=[\"Non-Fraud\", \"Fraud\"])\n",
        "plt.xlabel(\"Predicted\")\n",
        "plt.ylabel(\"Actual\")\n",
        "plt.title(\"Confusion Matrix\")\n",
        "plt.tight_layout()\n",
        "plt.show()\n",
        "\n",
        "# 分類報告\n",
        "print(\"分類報告：\")\n",
        "print(classification_report(y_test, y_pred))\n",
        "\n",
        "# AUC-ROC 分數\n",
        "auc = roc_auc_score(y_test, y_pred_prob)\n",
        "print(f\"AUC-ROC 分數：{auc:.4f}\")\n",
        "\n",
        "# 特徵重要性可視化\n",
        "xgb.plot_importance(xgb_model, importance_type='weight', max_num_features=10)\n",
        "plt.title(\"Top 10 Feature Importances\")\n",
        "plt.tight_layout()\n",
        "plt.show()\n",
        "\n",
        "# 結束 SparkSession\n",
        "spark.stop()\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "SHvjDBvX3WXJ"
      },
      "source": [
        "方法論：結合 XGBoost 與多種特徵篩選算法的模型優化\n",
        "在本研究中，我們結合了 XGBoost 的特徵重要性評估功能，以及貪婪算法、Dijkstra 算法與 Prim 算法的特徵選擇能力，構建了一個高效的特徵篩選與優化流程。這種方法旨在平衡模型性能與計算效率，適用於信用卡詐欺檢測這類高度不平衡的數據場景。\n",
        "\n",
        "第一階段：XGBoost 的特徵重要性評估\n",
        "XGBoost 作為一種基於樹的集成學習算法，不僅在分類任務中性能卓越，還能提供特徵的重要性分數。在此階段，我們基於 XGBoost 模型，對數據集中各個特徵的重要性進行排序，確定哪些特徵對分類結果影響最大。這些排名為後續特徵篩選提供了基礎。\n",
        "\n",
        "第二階段：多種特徵篩選算法的比較\n",
        "貪婪算法 (Greedy Algorithm)：\n",
        "貪婪算法從 XGBoost 的特徵重要性排名中逐步選擇特徵，每次新增一個特徵後評估其對模型性能的影響。當新增特徵不再帶來顯著提升時，算法停止，生成一個初步的高效特徵子集。此方法計算效率高，適合快速篩選場景。\n",
        "\n",
        "Dijkstra 算法：\n",
        "該算法基於圖的最短路徑思想，在特徵選擇過程中更智能地考慮特徵間的相互作用。與貪婪算法相比，Dijkstra 算法需要更多計算資源，但生成的特徵子集性能更接近於全特徵模型。\n",
        "\n",
        "Prim 算法：\n",
        "以最小生成樹的思想選擇特徵，Prim 算法每次選擇對模型性能提升最大的特徵，並考慮當前已選特徵與候選特徵的整體關係。該方法性能與 Dijkstra 算法相似，但計算成本更高，適合對精度要求極高的場景。\n",
        "\n",
        "第三階段：支持向量機 (SVM) 的應用\n",
        "在經過上述特徵篩選與優化後，我們使用支持向量機（SVM）模型進行分類。這些經過篩選的特徵子集顯著提升了 SVM 的分類性能，特別是在處理高度不平衡數據時，表現尤為突出。\n",
        "\n",
        "方法的綜合優勢\n",
        "高效性：貪婪算法能快速生成高效特徵子集，適合實時應用場景。\n",
        "精確性：Dijkstra 和 Prim 算法通過更智能的特徵選擇，生成性能更優的特徵子集。\n",
        "靈活性：多階段篩選方法能適應不同數據特性，無論是高維數據還是特徵冗餘問題，都能有效應對。\n",
        "結論\n",
        "我們的多階段特徵篩選與優化方法結合了 XGBoost 的特徵重要性評估和多種特徵選擇算法，不僅提升了模型性能，還有效平衡了計算資源的使用。在實際應用中，這種靈活的方法論可根據場景需求進一步調整，以適應不同的數據特性和性能要求。"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 10,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "id": "fz01dONd2UoB",
        "outputId": "ddec4d1b-52f0-4232-f715-a889a2e42fe0"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "特徵重要性排序：\n",
            "   Feature  Importance\n",
            "14     V14    0.284575\n",
            "10     V10    0.234879\n",
            "7       V7    0.073773\n",
            "12     V12    0.038514\n",
            "4       V4    0.030143\n",
            "17     V17    0.025719\n",
            "0     Time    0.019336\n",
            "1       V1    0.017758\n",
            "13     V13    0.015863\n",
            "29  Amount    0.015350\n",
            "18     V18    0.015340\n",
            "20     V20    0.015266\n",
            "21     V21    0.014923\n",
            "16     V16    0.014403\n",
            "26     V26    0.014264\n",
            "24     V24    0.013333\n",
            "9       V9    0.012765\n",
            "8       V8    0.012674\n",
            "3       V3    0.012621\n",
            "15     V15    0.012482\n",
            "27     V27    0.012104\n",
            "19     V19    0.011972\n",
            "28     V28    0.011932\n",
            "2       V2    0.011770\n",
            "6       V6    0.011479\n",
            "11     V11    0.010585\n",
            "23     V23    0.010058\n",
            "22     V22    0.009640\n",
            "5       V5    0.008792\n",
            "25     V25    0.007688\n"
          ]
        },
        {
          "data": {
            "image/png": "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",
            "text/plain": [
              "<Figure size 1000x600 with 1 Axes>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "選擇的特徵：['V14', 'V10', 'V12', 'V17', 'V1', 'V26', 'V3', 'V15', 'V19']\n",
            "優化後 SVM 的 AUC-ROC 分數：0.9557\n",
            "混淆矩陣：\n",
            "[[56862     2]\n",
            " [   83    15]]\n"
          ]
        },
        {
          "data": {
            "image/png": "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",
            "text/plain": [
              "<Figure size 800x600 with 2 Axes>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "分類報告：\n",
            "              precision    recall  f1-score   support\n",
            "\n",
            "           0       1.00      1.00      1.00     56864\n",
            "           1       0.88      0.15      0.26        98\n",
            "\n",
            "    accuracy                           1.00     56962\n",
            "   macro avg       0.94      0.58      0.63     56962\n",
            "weighted avg       1.00      1.00      1.00     56962\n",
            "\n"
          ]
        }
      ],
      "source": [
        "# 導入必要的庫\n",
        "from pyspark.sql import SparkSession\n",
        "from pyspark.sql.functions import col, when\n",
        "from pyspark.ml.feature import VectorAssembler\n",
        "from pyspark.ml.classification import LinearSVC\n",
        "from pyspark.ml.evaluation import BinaryClassificationEvaluator\n",
        "\n",
        "import pandas as pd\n",
        "import numpy as np\n",
        "import matplotlib.pyplot as plt\n",
        "import seaborn as sns\n",
        "import xgboost as xgb\n",
        "from sklearn.model_selection import train_test_split\n",
        "from sklearn.metrics import roc_auc_score, classification_report, confusion_matrix\n",
        "\n",
        "# 初始化 SparkSession\n",
        "spark = SparkSession.builder.appName(\"CreditCardFraudOptimization\").getOrCreate()\n",
        "\n",
        "# 讀取數據集\n",
        "file_path = \"/content/creditcard.csv\"  # 設置檔案路徑\n",
        "data = spark.read.csv(file_path, header=True, inferSchema=True)\n",
        "\n",
        "# 數據簡要查看\n",
        "data = data.withColumn(\"Class\", when(col(\"Class\") == 1, 1).otherwise(0))\n",
        "\n",
        "# 特徵組合\n",
        "assembler = VectorAssembler(inputCols=[col for col in data.columns if col != \"Class\"], outputCol=\"features\")\n",
        "data = assembler.transform(data)\n",
        "\n",
        "# 將 Spark DataFrame 轉換為 Pandas DataFrame\n",
        "pandas_data = data.select([col for col in data.columns if col != \"features\"]).toPandas()\n",
        "X = pandas_data.drop(columns=[\"Class\"])\n",
        "y = pandas_data[\"Class\"]\n",
        "\n",
        "# 分割數據集\n",
        "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42, stratify=y)\n",
        "\n",
        "# 使用 XGBoost 訓練模型\n",
        "xgb_model = xgb.XGBClassifier(eval_metric='auc', random_state=42)\n",
        "xgb_model.fit(X_train, y_train)\n",
        "\n",
        "# 提取特徵重要性\n",
        "feature_importances = xgb_model.feature_importances_\n",
        "feature_names = X.columns\n",
        "importance_df = pd.DataFrame({\"Feature\": feature_names, \"Importance\": feature_importances})\n",
        "importance_df = importance_df.sort_values(by=\"Importance\", ascending=False)\n",
        "\n",
        "# 顯示特徵重要性\n",
        "print(\"特徵重要性排序：\")\n",
        "print(importance_df)\n",
        "\n",
        "# 可視化特徵重要性\n",
        "plt.figure(figsize=(10, 6))\n",
        "sns.barplot(x=\"Importance\", y=\"Feature\", data=importance_df)\n",
        "plt.title(\"Feature Importance (XGBoost)\")\n",
        "plt.tight_layout()\n",
        "plt.show()\n",
        "\n",
        "# 特徵篩選 - 貪婪算法\n",
        "selected_features = []\n",
        "current_features = []\n",
        "current_auc = 0\n",
        "\n",
        "for feature in importance_df[\"Feature\"]:\n",
        "    current_features.append(feature)\n",
        "    X_train_subset = X_train[current_features]\n",
        "    X_test_subset = X_test[current_features]\n",
        "\n",
        "    temp_model = xgb.XGBClassifier(eval_metric='auc', random_state=42)\n",
        "    temp_model.fit(X_train_subset, y_train)\n",
        "    y_pred_prob = temp_model.predict_proba(X_test_subset)[:, 1]\n",
        "    temp_auc = roc_auc_score(y_test, y_pred_prob)\n",
        "\n",
        "    if temp_auc > current_auc:\n",
        "        current_auc = temp_auc\n",
        "        selected_features = list(current_features)\n",
        "    else:\n",
        "        current_features.pop()  # 如果性能下降，移除該特徵\n",
        "\n",
        "print(f\"選擇的特徵：{selected_features}\")\n",
        "\n",
        "# 將 Pandas DataFrame 轉換回 Spark DataFrame\n",
        "X_train_final = X_train[selected_features].copy()\n",
        "X_train_final.loc[:, \"Class\"] = y_train.values\n",
        "X_test_final = X_test[selected_features].copy()\n",
        "X_test_final.loc[:, \"Class\"] = y_test.values\n",
        "\n",
        "train_data = spark.createDataFrame(X_train_final)\n",
        "test_data = spark.createDataFrame(X_test_final)\n",
        "\n",
        "# 使用 VectorAssembler 將選擇的特徵組合\n",
        "assembler = VectorAssembler(inputCols=selected_features, outputCol=\"features\")\n",
        "train_data = assembler.transform(train_data)\n",
        "test_data = assembler.transform(test_data)\n",
        "\n",
        "# 訓練 SVM 模型\n",
        "svm = LinearSVC(featuresCol=\"features\", labelCol=\"Class\", maxIter=10, regParam=0.1)\n",
        "svm_model = svm.fit(train_data)\n",
        "\n",
        "# 預測\n",
        "predictions = svm_model.transform(test_data)\n",
        "\n",
        "# 評估性能\n",
        "evaluator = BinaryClassificationEvaluator(labelCol=\"Class\", rawPredictionCol=\"rawPrediction\", metricName=\"areaUnderROC\")\n",
        "svm_auc = evaluator.evaluate(predictions)\n",
        "print(f\"優化後 SVM 的 AUC-ROC 分數：{svm_auc:.4f}\")\n",
        "\n",
        "# 混淆矩陣\n",
        "predictions_pd = predictions.select(\"Class\", \"prediction\").toPandas()\n",
        "conf_matrix = confusion_matrix(predictions_pd[\"Class\"], predictions_pd[\"prediction\"])\n",
        "\n",
        "print(\"混淆矩陣：\")\n",
        "print(conf_matrix)\n",
        "\n",
        "# 混淆矩陣可視化\n",
        "plt.figure(figsize=(8, 6))\n",
        "sns.heatmap(conf_matrix, annot=True, fmt=\"d\", cmap=\"Blues\", xticklabels=[\"Non-Fraud\", \"Fraud\"], yticklabels=[\"Non-Fraud\", \"Fraud\"])\n",
        "plt.xlabel(\"Predicted\")\n",
        "plt.ylabel(\"Actual\")\n",
        "plt.title(\"Confusion Matrix\")\n",
        "plt.tight_layout()\n",
        "plt.show()\n",
        "\n",
        "# 分類報告\n",
        "print(\"分類報告：\")\n",
        "print(classification_report(predictions_pd[\"Class\"], predictions_pd[\"prediction\"]))\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 11,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "id": "pVUtVYD-F3p9",
        "outputId": "d8b2bdcf-5c01-4a54-b938-0b481bb78549"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "特徵重要性排序：\n",
            "   Feature  Importance\n",
            "14     V14    0.284575\n",
            "10     V10    0.234879\n",
            "7       V7    0.073773\n",
            "12     V12    0.038514\n",
            "4       V4    0.030143\n",
            "17     V17    0.025719\n",
            "0     Time    0.019336\n",
            "1       V1    0.017758\n",
            "13     V13    0.015863\n",
            "29  Amount    0.015350\n",
            "18     V18    0.015340\n",
            "20     V20    0.015266\n",
            "21     V21    0.014923\n",
            "16     V16    0.014403\n",
            "26     V26    0.014264\n",
            "24     V24    0.013333\n",
            "9       V9    0.012765\n",
            "8       V8    0.012674\n",
            "3       V3    0.012621\n",
            "15     V15    0.012482\n",
            "27     V27    0.012104\n",
            "19     V19    0.011972\n",
            "28     V28    0.011932\n",
            "2       V2    0.011770\n",
            "6       V6    0.011479\n",
            "11     V11    0.010585\n",
            "23     V23    0.010058\n",
            "22     V22    0.009640\n",
            "5       V5    0.008792\n",
            "25     V25    0.007688\n"
          ]
        },
        {
          "data": {
            "image/png": "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",
            "text/plain": [
              "<Figure size 1000x600 with 1 Axes>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "選擇的特徵：['V14', 'V12', 'V5', 'V25', 'Amount', 'V20', 'V19', 'V22', 'V28', 'V23', 'V2', 'V11', 'V6', 'V15', 'V27', 'V8', 'V3', 'V9', 'V26', 'V24', 'V21', 'V16', 'V18', 'V4', 'Time', 'V1', 'V13', 'V17', 'V10', 'V7']\n",
            "優化後 SVM 的 AUC-ROC 分數：0.9786\n",
            "混淆矩陣：\n",
            "[[56851    13]\n",
            " [   40    58]]\n"
          ]
        },
        {
          "data": {
            "image/png": "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",
            "text/plain": [
              "<Figure size 800x600 with 2 Axes>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "分類報告：\n",
            "              precision    recall  f1-score   support\n",
            "\n",
            "           0       1.00      1.00      1.00     56864\n",
            "           1       0.82      0.59      0.69        98\n",
            "\n",
            "    accuracy                           1.00     56962\n",
            "   macro avg       0.91      0.80      0.84     56962\n",
            "weighted avg       1.00      1.00      1.00     56962\n",
            "\n"
          ]
        }
      ],
      "source": [
        "# 導入必要的庫\n",
        "from pyspark.sql import SparkSession\n",
        "from pyspark.sql.functions import col, when\n",
        "from pyspark.ml.feature import VectorAssembler\n",
        "from pyspark.ml.classification import LinearSVC\n",
        "from pyspark.ml.evaluation import BinaryClassificationEvaluator\n",
        "\n",
        "import pandas as pd\n",
        "import numpy as np\n",
        "import matplotlib.pyplot as plt\n",
        "import seaborn as sns\n",
        "import xgboost as xgb\n",
        "from sklearn.model_selection import train_test_split\n",
        "from sklearn.metrics import roc_auc_score, classification_report, confusion_matrix\n",
        "\n",
        "# 初始化 SparkSession\n",
        "spark = SparkSession.builder.appName(\"CreditCardFraudOptimization\").getOrCreate()\n",
        "\n",
        "# 讀取數據集\n",
        "file_path = \"/content/creditcard.csv\"  # 設置檔案路徑\n",
        "data = spark.read.csv(file_path, header=True, inferSchema=True)\n",
        "\n",
        "# 數據簡要查看\n",
        "data = data.withColumn(\"Class\", when(col(\"Class\") == 1, 1).otherwise(0))\n",
        "\n",
        "# 特徵組合\n",
        "assembler = VectorAssembler(inputCols=[col for col in data.columns if col != \"Class\"], outputCol=\"features\")\n",
        "data = assembler.transform(data)\n",
        "\n",
        "# 將 Spark DataFrame 轉換為 Pandas DataFrame\n",
        "pandas_data = data.select([col for col in data.columns if col != \"features\"]).toPandas()\n",
        "X = pandas_data.drop(columns=[\"Class\"])\n",
        "y = pandas_data[\"Class\"]\n",
        "\n",
        "# 分割數據集\n",
        "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42, stratify=y)\n",
        "\n",
        "# 使用 XGBoost 訓練模型\n",
        "xgb_model = xgb.XGBClassifier(eval_metric='auc', random_state=42)\n",
        "xgb_model.fit(X_train, y_train)\n",
        "\n",
        "# 提取特徵重要性\n",
        "feature_importances = xgb_model.feature_importances_\n",
        "feature_names = X.columns\n",
        "importance_df = pd.DataFrame({\"Feature\": feature_names, \"Importance\": feature_importances})\n",
        "importance_df = importance_df.sort_values(by=\"Importance\", ascending=False)\n",
        "\n",
        "# 顯示特徵重要性\n",
        "print(\"特徵重要性排序：\")\n",
        "print(importance_df)\n",
        "\n",
        "# 可視化特徵重要性\n",
        "plt.figure(figsize=(10, 6))\n",
        "sns.barplot(x=\"Importance\", y=\"Feature\", data=importance_df)\n",
        "plt.title(\"Feature Importance (XGBoost)\")\n",
        "plt.tight_layout()\n",
        "plt.show()\n",
        "\n",
        "# 特徵篩選 - 使用 Dijkstra 演算法進行篩選\n",
        "def dijkstra_feature_selection(features, X_train, X_test, y_train, y_test):\n",
        "    n = len(features)\n",
        "    best_features = []\n",
        "    visited = set()\n",
        "    distances = {feature: float('inf') for feature in features}\n",
        "    distances[features[0]] = 0  # 假設第一個特徵為起點\n",
        "\n",
        "    while len(visited) < n:\n",
        "        # 找到尚未訪問的特徵中距離最小的\n",
        "        current_feature = min((f for f in features if f not in visited), key=lambda f: distances[f])\n",
        "        visited.add(current_feature)\n",
        "\n",
        "        # 更新相鄰特徵的距離\n",
        "        current_index = features.index(current_feature)\n",
        "        for neighbor_index in range(current_index + 1, n):\n",
        "            neighbor = features[neighbor_index]\n",
        "            if neighbor not in visited:\n",
        "                X_train_subset = X_train[best_features + [neighbor]]\n",
        "                X_test_subset = X_test[best_features + [neighbor]]\n",
        "\n",
        "                temp_model = xgb.XGBClassifier(eval_metric='auc', random_state=42)\n",
        "                temp_model.fit(X_train_subset, y_train)\n",
        "                y_pred_prob = temp_model.predict_proba(X_test_subset)[:, 1]\n",
        "                temp_auc = roc_auc_score(y_test, y_pred_prob)\n",
        "\n",
        "                # 使用 AUC 值作為距離的反向測度\n",
        "                distance = 1 - temp_auc\n",
        "                if distance < distances[neighbor]:\n",
        "                    distances[neighbor] = distance\n",
        "\n",
        "        best_features.append(current_feature)\n",
        "\n",
        "    return best_features\n",
        "\n",
        "selected_features = dijkstra_feature_selection(list(importance_df[\"Feature\"]), X_train, X_test, y_train, y_test)\n",
        "print(f\"選擇的特徵：{selected_features}\")\n",
        "\n",
        "# 將 Pandas DataFrame 轉換回 Spark DataFrame\n",
        "X_train_final = X_train[selected_features].copy()\n",
        "X_train_final.loc[:, \"Class\"] = y_train.values\n",
        "X_test_final = X_test[selected_features].copy()\n",
        "X_test_final.loc[:, \"Class\"] = y_test.values\n",
        "\n",
        "train_data = spark.createDataFrame(X_train_final)\n",
        "test_data = spark.createDataFrame(X_test_final)\n",
        "\n",
        "# 使用 VectorAssembler 將選擇的特徵組合\n",
        "assembler = VectorAssembler(inputCols=selected_features, outputCol=\"features\")\n",
        "train_data = assembler.transform(train_data)\n",
        "test_data = assembler.transform(test_data)\n",
        "\n",
        "# 訓練 SVM 模型\n",
        "svm = LinearSVC(featuresCol=\"features\", labelCol=\"Class\", maxIter=10, regParam=0.1)\n",
        "svm_model = svm.fit(train_data)\n",
        "\n",
        "# 預測\n",
        "predictions = svm_model.transform(test_data)\n",
        "\n",
        "# 評估性能\n",
        "evaluator = BinaryClassificationEvaluator(labelCol=\"Class\", rawPredictionCol=\"rawPrediction\", metricName=\"areaUnderROC\")\n",
        "svm_auc = evaluator.evaluate(predictions)\n",
        "print(f\"優化後 SVM 的 AUC-ROC 分數：{svm_auc:.4f}\")\n",
        "\n",
        "# 混淆矩陣\n",
        "predictions_pd = predictions.select(\"Class\", \"prediction\").toPandas()\n",
        "conf_matrix = confusion_matrix(predictions_pd[\"Class\"], predictions_pd[\"prediction\"])\n",
        "\n",
        "print(\"混淆矩陣：\")\n",
        "print(conf_matrix)\n",
        "\n",
        "# 混淆矩陣可視化\n",
        "plt.figure(figsize=(8, 6))\n",
        "sns.heatmap(conf_matrix, annot=True, fmt=\"d\", cmap=\"Blues\", xticklabels=[\"Non-Fraud\", \"Fraud\"], yticklabels=[\"Non-Fraud\", \"Fraud\"])\n",
        "plt.xlabel(\"Predicted\")\n",
        "plt.ylabel(\"Actual\")\n",
        "plt.title(\"Confusion Matrix\")\n",
        "plt.tight_layout()\n",
        "plt.show()\n",
        "\n",
        "# 分類報告\n",
        "print(\"分類報告：\")\n",
        "print(classification_report(predictions_pd[\"Class\"], predictions_pd[\"prediction\"]))\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 12,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "id": "IohdROhMG42z",
        "outputId": "f9707ffb-85d3-4b88-a753-ec35bb9809a5"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "特徵重要性排序：\n",
            "   Feature  Importance\n",
            "14     V14    0.284575\n",
            "10     V10    0.234879\n",
            "7       V7    0.073773\n",
            "12     V12    0.038514\n",
            "4       V4    0.030143\n",
            "17     V17    0.025719\n",
            "0     Time    0.019336\n",
            "1       V1    0.017758\n",
            "13     V13    0.015863\n",
            "29  Amount    0.015350\n",
            "18     V18    0.015340\n",
            "20     V20    0.015266\n",
            "21     V21    0.014923\n",
            "16     V16    0.014403\n",
            "26     V26    0.014264\n",
            "24     V24    0.013333\n",
            "9       V9    0.012765\n",
            "8       V8    0.012674\n",
            "3       V3    0.012621\n",
            "15     V15    0.012482\n",
            "27     V27    0.012104\n",
            "19     V19    0.011972\n",
            "28     V28    0.011932\n",
            "2       V2    0.011770\n",
            "6       V6    0.011479\n",
            "11     V11    0.010585\n",
            "23     V23    0.010058\n",
            "22     V22    0.009640\n",
            "5       V5    0.008792\n",
            "25     V25    0.007688\n"
          ]
        },
        {
          "data": {
            "image/png": "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",
            "text/plain": [
              "<Figure size 1000x600 with 1 Axes>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "選擇的特徵：['V14', 'V5', 'Amount', 'V12', 'V19', 'V22', 'V28', 'V18', 'V21', 'V15', 'V2', 'Time', 'V20', 'V6', 'V9', 'V24', 'V23', 'V13', 'V25', 'V8', 'V3', 'V7', 'V17', 'V16', 'V10', 'V1', 'V27', 'V11', 'V26', 'V4']\n",
            "優化後 SVM 的 AUC-ROC 分數：0.9786\n",
            "混淆矩陣：\n",
            "[[56851    13]\n",
            " [   40    58]]\n"
          ]
        },
        {
          "data": {
            "image/png": "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",
            "text/plain": [
              "<Figure size 800x600 with 2 Axes>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "分類報告：\n",
            "              precision    recall  f1-score   support\n",
            "\n",
            "           0       1.00      1.00      1.00     56864\n",
            "           1       0.82      0.59      0.69        98\n",
            "\n",
            "    accuracy                           1.00     56962\n",
            "   macro avg       0.91      0.80      0.84     56962\n",
            "weighted avg       1.00      1.00      1.00     56962\n",
            "\n"
          ]
        }
      ],
      "source": [
        "# 導入必要的庫\n",
        "from pyspark.sql import SparkSession\n",
        "from pyspark.sql.functions import col, when\n",
        "from pyspark.ml.feature import VectorAssembler\n",
        "from pyspark.ml.classification import LinearSVC\n",
        "from pyspark.ml.evaluation import BinaryClassificationEvaluator\n",
        "\n",
        "import pandas as pd\n",
        "import numpy as np\n",
        "import matplotlib.pyplot as plt\n",
        "import seaborn as sns\n",
        "import xgboost as xgb\n",
        "from sklearn.model_selection import train_test_split\n",
        "from sklearn.metrics import roc_auc_score, classification_report, confusion_matrix\n",
        "\n",
        "# 初始化 SparkSession\n",
        "spark = SparkSession.builder.appName(\"CreditCardFraudOptimization\").getOrCreate()\n",
        "\n",
        "# 讀取數據集\n",
        "file_path = \"/content/creditcard.csv\"  # 設置檔案路徑\n",
        "data = spark.read.csv(file_path, header=True, inferSchema=True)\n",
        "\n",
        "# 數據簡要查看\n",
        "data = data.withColumn(\"Class\", when(col(\"Class\") == 1, 1).otherwise(0))\n",
        "\n",
        "# 特徵組合\n",
        "assembler = VectorAssembler(inputCols=[col for col in data.columns if col != \"Class\"], outputCol=\"features\")\n",
        "data = assembler.transform(data)\n",
        "\n",
        "# 將 Spark DataFrame 轉換為 Pandas DataFrame\n",
        "pandas_data = data.select([col for col in data.columns if col != \"features\"]).toPandas()\n",
        "X = pandas_data.drop(columns=[\"Class\"])\n",
        "y = pandas_data[\"Class\"]\n",
        "\n",
        "# 分割數據集\n",
        "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42, stratify=y)\n",
        "\n",
        "# 使用 XGBoost 訓練模型\n",
        "xgb_model = xgb.XGBClassifier(eval_metric='auc', random_state=42)\n",
        "xgb_model.fit(X_train, y_train)\n",
        "\n",
        "# 提取特徵重要性\n",
        "feature_importances = xgb_model.feature_importances_\n",
        "feature_names = X.columns\n",
        "importance_df = pd.DataFrame({\"Feature\": feature_names, \"Importance\": feature_importances})\n",
        "importance_df = importance_df.sort_values(by=\"Importance\", ascending=False)\n",
        "\n",
        "# 顯示特徵重要性\n",
        "print(\"特徵重要性排序：\")\n",
        "print(importance_df)\n",
        "\n",
        "# 可視化特徵重要性\n",
        "plt.figure(figsize=(10, 6))\n",
        "sns.barplot(x=\"Importance\", y=\"Feature\", data=importance_df)\n",
        "plt.title(\"Feature Importance (XGBoost)\")\n",
        "plt.tight_layout()\n",
        "plt.show()\n",
        "\n",
        "# 特徵篩選 - 使用 Prim 演算法進行篩選\n",
        "def prim_feature_selection(features, X_train, X_test, y_train, y_test):\n",
        "    n = len(features)\n",
        "    selected_features = [features[0]]  # 從第一個特徵開始\n",
        "    available_features = set(features[1:])\n",
        "\n",
        "    while available_features:\n",
        "        best_feature = None\n",
        "        best_auc = 0\n",
        "\n",
        "        for feature in available_features:\n",
        "            temp_features = selected_features + [feature]\n",
        "            X_train_subset = X_train[temp_features]\n",
        "            X_test_subset = X_test[temp_features]\n",
        "\n",
        "            temp_model = xgb.XGBClassifier(eval_metric='auc', random_state=42)\n",
        "            temp_model.fit(X_train_subset, y_train)\n",
        "            y_pred_prob = temp_model.predict_proba(X_test_subset)[:, 1]\n",
        "            temp_auc = roc_auc_score(y_test, y_pred_prob)\n",
        "\n",
        "            if temp_auc > best_auc:\n",
        "                best_auc = temp_auc\n",
        "                best_feature = feature\n",
        "\n",
        "        if best_feature:\n",
        "            selected_features.append(best_feature)\n",
        "            available_features.remove(best_feature)\n",
        "        else:\n",
        "            break\n",
        "\n",
        "    return selected_features\n",
        "\n",
        "selected_features = prim_feature_selection(list(importance_df[\"Feature\"]), X_train, X_test, y_train, y_test)\n",
        "print(f\"選擇的特徵：{selected_features}\")\n",
        "\n",
        "# 將 Pandas DataFrame 轉換回 Spark DataFrame\n",
        "X_train_final = X_train[selected_features].copy()\n",
        "X_train_final.loc[:, \"Class\"] = y_train.values\n",
        "X_test_final = X_test[selected_features].copy()\n",
        "X_test_final.loc[:, \"Class\"] = y_test.values\n",
        "\n",
        "train_data = spark.createDataFrame(X_train_final)\n",
        "test_data = spark.createDataFrame(X_test_final)\n",
        "\n",
        "# 使用 VectorAssembler 將選擇的特徵組合\n",
        "assembler = VectorAssembler(inputCols=selected_features, outputCol=\"features\")\n",
        "train_data = assembler.transform(train_data)\n",
        "test_data = assembler.transform(test_data)\n",
        "\n",
        "# 訓練 SVM 模型\n",
        "svm = LinearSVC(featuresCol=\"features\", labelCol=\"Class\", maxIter=10, regParam=0.1)\n",
        "svm_model = svm.fit(train_data)\n",
        "\n",
        "# 預測\n",
        "predictions = svm_model.transform(test_data)\n",
        "\n",
        "# 評估性能\n",
        "evaluator = BinaryClassificationEvaluator(labelCol=\"Class\", rawPredictionCol=\"rawPrediction\", metricName=\"areaUnderROC\")\n",
        "svm_auc = evaluator.evaluate(predictions)\n",
        "print(f\"優化後 SVM 的 AUC-ROC 分數：{svm_auc:.4f}\")\n",
        "\n",
        "# 混淆矩陣\n",
        "predictions_pd = predictions.select(\"Class\", \"prediction\").toPandas()\n",
        "conf_matrix = confusion_matrix(predictions_pd[\"Class\"], predictions_pd[\"prediction\"])\n",
        "\n",
        "print(\"混淆矩陣：\")\n",
        "print(conf_matrix)\n",
        "\n",
        "# 混淆矩陣可視化\n",
        "plt.figure(figsize=(8, 6))\n",
        "sns.heatmap(conf_matrix, annot=True, fmt=\"d\", cmap=\"Blues\", xticklabels=[\"Non-Fraud\", \"Fraud\"], yticklabels=[\"Non-Fraud\", \"Fraud\"])\n",
        "plt.xlabel(\"Predicted\")\n",
        "plt.ylabel(\"Actual\")\n",
        "plt.title(\"Confusion Matrix\")\n",
        "plt.tight_layout()\n",
        "plt.show()\n",
        "\n",
        "# 分類報告\n",
        "print(\"分類報告：\")\n",
        "print(classification_report(predictions_pd[\"Class\"], predictions_pd[\"prediction\"]))\n"
      ]
    }
  ],
  "metadata": {
    "colab": {
      "provenance": []
    },
    "kernelspec": {
      "display_name": "Python 3",
      "name": "python3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "nbformat": 4,
  "nbformat_minor": 0
}