使用Spark MLlib的逻辑回归(LogisticRegression)进行用户分类预测识别

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

// 2.解析数据，构建数据集
val parsedTrainData = trainData.map { line =>
val parts= line.split("\\|")
val label = toInt(parts(1)) //第二列是标签
val features = Vectors.dense(parts.slice(6,parts.length-1).map(_.toDouble)) //第7到最后一列是属性，需要转换为Doube类型
LabeledPoint(label, features) //构建LabelPoint格式,第一列是标签列，后面是属性向量
}.cache()

// 3.将数据集随机分为两份，一份是训练集，一份是测试集
val splits = parsedTrainData.randomSplit(Array(0.7, 0.3), seed = 11L)
val training = splits(0)
val testing = splits(1)

// 4.新建逻辑回归模型，并设置训练参数
//    val model = new LogisticRegressionWithLBFGS().setNumClasses(2)
//    model.optimizer.setNumIterations(500).setUpdater(new SimpleUpdater())

//可以选择LogisticRegressionWithLBFGS，也可以选择LogisticRegressionWithSGD，LogisticRegressionWithLBFGS是优化方法
val model = new LogisticRegressionWithSGD()  //建立模型
model.optimizer.setNumIterations(500).setUpdater(new SimpleUpdater()).setStepSize(0.001).setMiniBatchFraction(0.02) //模型参数

val trained = model.run(training)  //使用训练集训练模型

// 5.测试样本进行预测
val prediction = trained.predict(testing.map(_.features)) //使用测试数据属性进行预测

val predictionAndLabels = prediction.zip(testing.map(_.label)) //获取预测标签

// 6.测量预测效果
val metrics = new BinaryClassificationMetrics(predictionAndLabels)

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

// 将标签转换为0和1
def toInt(s: String): Int = {
if (s == "m") 1 else  0
}

}