對多組資料進行model的training,然後再利用model來predict具體的值
。過程中有輸出model的權重
公式：f(x)=a1X1+a2X2+a3X3+……

2.程式碼：

package org.apache.spark.mllib.learning.regression

import java.text.SimpleDateFormat
import java.util.Date

import org.apache.log4j.{Level, Logger}
import org.apache.spark.mllib.linalg.Vectors
import org.apache
 
.spark.mllib.regression.{LabeledPoint, LinearRegressionWithSGD}
import org.apache.spark.{SparkConf, SparkContext}

import scala.Array.canBuildFrom

object LinearRegression {
  def main(args: Array[String]): Unit = {
    // 遮蔽不必要的日誌顯示終端上
    Logger.getLogger("org.apache.spark").setLevel(Level.ERROR)
    Logger.getLogger
 
("org.eclipse.jetty.server").setLevel(Level.OFF)

    // 設定執行環境
    val conf = new SparkConf().setAppName(this.getClass().getSimpleName().filter(!_.equals('$'))).setMaster("local[4]")
    val sc = new SparkContext(conf)

    // Load and parse the data
    val data = sc.textFile("file/data/mllib/input/ridge-data/lpsa.data"
 
,1)
    //如果讀入不加1，會產生兩個檔案，應該是預設生成了兩個partition
    val parsedData = data.map { line =>
      val parts = line.split(',')
      LabeledPoint(parts(0).toDouble, Vectors.dense(parts(1).split(' ').map(_.toDouble)))
    }

    // Building the model
    //建立model的資料和predict的資料沒有分開
    val numIterations = 100
    val model = LinearRegressionWithSGD.train(parsedData, numIterations)
    //    for(i<-parsedData) println(i.label+":"+i.features);
    // Evaluate model on training examples and compute training error
    val valuesAndPreds = parsedData.map { point =>
      val prediction = model.predict(point.features)
      (point.label, prediction)
    }
    //print model.weights
    var weifhts=model.weights
    println("model.weights"+weifhts)

    //save as file
    val iString = new SimpleDateFormat("yyyyMMddHHmmssSSS").format(new Date())
    val path = "file/data/mllib/output/LinearRegression/" + iString + "/result"
    valuesAndPreds.saveAsTextFile(path)
    val MSE = valuesAndPreds.map { case (v, p) => math.pow((v - p), 2) }.reduce(_ + _) / valuesAndPreds.count
    println("training Mean Squared Error = " + MSE)

    sc.stop()
  }
}

資料請見github或者spark原始碼

3.結果：

model.weights[0.5808575763272221,0.18930001482946976,0.2803086929991066,0.1110834181777876,0.4010473965597895,-0.5603061626684255,-0.5804740464000981,0.8742741176970946]
training Mean Squared Error = 6.207597210613579