import org.apache.spark.SparkContext
import org.apache.spark.SparkConf
import org.apache.spark.mllib.classification.{LogisticRegressionWithLBFGS, LogisticRegressionWithSGD}
import org.apache.spark.mllib.regression.LabeledPoint
import org.apache.spark.mllib.linalg.Vectors
import org.apache.spark.mllib.evaluation.BinaryClassificationMetrics
import
 
 org.apache.spark.mllib.optimization._

/**
  * Created by simon on 2017/5/8.
  */
object genderClassificationWithLogisticRegression {
  def main(args: Array[String]): Unit = {
    val conf = new SparkConf()
    conf.setAppName("genderClassification").setMaster("local[2]")
    val sc = new SparkContext(conf)
    // 1.讀取資料
 

    val trainData = sc.textFile("file:\\E:\\test.csv")

    // 2.解析資料，構建資料集
    val parsedTrainData = trainData.map { line =>
      val parts= line.split("\\|")
      val label = toInt(parts(1)) //第二列是標籤
      val features = Vectors.dense(parts.slice(6,parts.length-1).map(_.toDouble)) //第7到最後一列是屬性，需要轉換為Doube型別
 

      LabeledPoint(label, features) //構建LabelPoint格式,第一列是標籤列，後面是屬性向量
    }.cache()

    // 3.將資料集隨機分為兩份，一份是訓練集，一份是測試集
    val splits = parsedTrainData.randomSplit(Array(0.7, 0.3), seed = 11L)
    val training = splits(0)
    val testing = splits(1)

    // 4.新建邏輯迴歸模型，並設定訓練引數
//    val model = new LogisticRegressionWithLBFGS().setNumClasses(2)
//    model.optimizer.setNumIterations(500).setUpdater(new SimpleUpdater())

//可以選擇LogisticRegressionWithLBFGS，也可以選擇LogisticRegressionWithSGD，LogisticRegressionWithLBFGS是優化方法
    val model = new LogisticRegressionWithSGD()  //建立模型
    model.optimizer.setNumIterations(500).setUpdater(new SimpleUpdater()).setStepSize(0.001).setMiniBatchFraction(0.02) //模型引數

    val trained = model.run(training)  //使用訓練集訓練模型

    // 5.測試樣本進行預測
    val prediction = trained.predict(testing.map(_.features)) //使用測試資料屬性進行預測

    val predictionAndLabels = prediction.zip(testing.map(_.label)) //獲取預測標籤

    // 6.測量預測效果
    val metrics = new BinaryClassificationMetrics(predictionAndLabels)

    // 7.看看AUROC結果
    val auROC = metrics.areaUnderROC
    println("Area under ROC = " + auROC)
  }

  // 將標籤轉換為0和1
  def toInt(s: String): Int = {
    if (s == "m") 1 else  0
  }

}