如何为 setInputCol() 提供多列

Question

我对 Spark 机器学习非常陌生，我想将多个列传递给功能，在下面的代码中，我只将日期列传递给功能，但现在我想将用户 ID 和日期列传递给功能。我尝试使用 Vector 但它只支持 Double 数据类型但在我的情况下我有 Int 和 String

如果有人提供任何可以满足我要求的建议/解决方案或任何代码示例，我将不胜感激

代码：

 case class LabeledDocument(Userid: Double, Date: String, label: Double)
 val training = spark.read.option("inferSchema", true).csv("/root/Predictiondata3.csv").toDF("Userid","Date","label").toDF().as[LabeledDocument]
 import scala.beans.BeanInfo
 import org.apache.spark.{SparkConf, SparkContext}
 import org.apache.spark.ml.Pipeline
 import org.apache.spark.ml.classification.LogisticRegression
 import org.apache.spark.ml.feature.{HashingTF, Tokenizer}
 import org.apache.spark.mllib.linalg.Vector
 import org.apache.spark.sql.{Row, SQLContext}
 val tokenizer = new Tokenizer().setInputCol("Date").setOutputCol("words")
 val hashingTF = new HashingTF().setNumFeatures(1000).setInputCol(tokenizer.getOutputCol).setOutputCol("features")
 import org.apache.spark.ml.regression.LinearRegression
 val lr = new LinearRegression().setMaxIter(100).setRegParam(0.001).setElasticNetParam(0.0001)
 val pipeline = new Pipeline().setStages(Array(tokenizer, hashingTF, lr))
 val model = pipeline.fit(training.toDF())
 case class Document(Userid: Integer, Date: String)
 val test = sc.parallelize(Seq(Document(4, "04-Jan-18"),Document(5, "01-Jan-17"),Document(2, "03-Jan-17")))
 model.transform(test.toDF()).show()

输入数据与列

Userid,Date,SwipeIntime
1, 1-Jan-2017,9.30
1, 2-Jan-2017,9.35
1, 3-Jan-2017,9.45
1, 4-Jan-2017,9.26
2, 1-Jan-2017,9.37
2, 2-Jan-2017,9.35
2, 3-Jan-2017,9.45
2, 4-Jan-2017,9.46  

Answer 1

我得到了我能够做到的解决方案。

 import scala.beans.BeanInfo
 import org.apache.spark.{SparkConf, SparkContext}
 import org.apache.spark.ml.Pipeline
 import org.apache.spark.ml.classification.LogisticRegression
 import org.apache.spark.ml.feature.{HashingTF, Tokenizer}
 import org.apache.spark.mllib.linalg.Vector
 import org.apache.spark.sql.{Row, SQLContext}
 import org.apache.spark.mllib.linalg.Vectors
 import org.apache.spark.ml.attribute.NominalAttribute
 import org.apache.spark.sql.Row
 import org.apache.spark.sql.types.{StructType,StructField,StringType}
 case class LabeledDocument(Userid: Double, Date: String, label: Double)
 val trainingData = spark.read.option("inferSchema", true).csv("/root/Predictiondata10.csv").toDF("Userid","Date","label").toDF().as[LabeledDocument]
 import org.apache.spark.ml.feature.StringIndexer
 import org.apache.spark.ml.feature.VectorAssembler
 val DateIndexer = new StringIndexer().setInputCol("Date").setOutputCol("DateCat")
 val indexed = DateIndexer.fit(trainingData).transform(trainingData)
 val assembler = new VectorAssembler().setInputCols(Array("DateCat", "Userid")).setOutputCol("rawfeatures")
 val output = assembler.transform(indexed)
 val rows = output.select("Userid","Date","label","DateCat","rawfeatures").collect()
 val asTuple=rows.map(a=>(a.getInt(0),a.getString(1),a.getDouble(2),a.getDouble(3),a(4).toString()))
 val r2 = sc.parallelize(asTuple).toDF("Userid","Date","label","DateCat","rawfeatures")
 val Array(training, testData) = r2.randomSplit(Array(0.7, 0.3))
 import org.apache.spark.ml.feature.{HashingTF, Tokenizer}
 val tokenizer = new Tokenizer().setInputCol("rawfeatures").setOutputCol("words")
 val hashingTF = new HashingTF().setNumFeatures(1000).setInputCol(tokenizer.getOutputCol).setOutputCol("features")
 import org.apache.spark.ml.regression.LinearRegression
 val lr = new LinearRegression().setMaxIter(100).setRegParam(0.001).setElasticNetParam(0.0001)
 val pipeline = new Pipeline().setStages(Array(tokenizer, hashingTF, lr))
 val model = pipeline.fit(training.toDF())
 model.transform(testData.toDF()).show()