Spark--LinearRegressionWithSGD

测试数据：链接: https://pan.baidu.com/s/1i7owaXJ 密码: 4wqg

代码记录：

import org.apache.spark.mllib.linalg.Vectors
import org.apache.spark.mllib.regression.{LabeledPoint, LinearRegressionWithSGD}
import org.apache.spark.sql.SparkSession

/**
  * Created by root on 1/12/18.
  */
object LinearRegressionWithSGDTest {
  def main(args: Array[String]): Unit = {
    val path = "/home/enche/data/lpsa.data"
    val spark = SparkSession.builder().appName("LinearRegressionWithSGD").master("local").getOrCreate()
    val sc = spark.sparkContext
    val data = sc.textFile(path)
//将数据转成Rdd[LabelPoint]格式
    val traindata = data.map(
      line=>LabeledPoint(
        line.split(",")(0).toDouble, Vectors.dense(line.split(",")(1).split(" ").map(_.toDouble))
      )
    )
//train(Rdd[LabelPoint],最大迭代数，学习率，每批次占总数的百分比)
    val model = LinearRegressionWithSGD.train(traindata,100,0.2,0.1)
    print(model.weights)
  }
}