I was using spark ML pipeline to set up classification models on really wide table. This means that I have to automatically generate all the code that deals with columns instead of literately typing each of them. I am pretty much a beginner on scala and spark. I was stuck at the VectorAssembler() part when I was trying to do something like following:
val featureHeaders = featureHeader.collect.mkString(" ")
//convert the header RDD into a string
val featureArray = featureHeaders.split(",").toArray
val quote = "\""
val featureSIArray = featureArray.map(x => (s"$quote$x$quote"))
//count the element in headers
val featureHeader_cnt = featureHeaders.split(",").toList.length
// Fit on whole dataset to include all labels in index.
import org.apache.spark.ml.feature.StringIndexer
val labelIndexer = new StringIndexer().
setInputCol("target").
setOutputCol("indexedLabel")
val featureAssembler = new VectorAssembler().
setInputCols(featureSIArray).
setOutputCol("features")
val convpipeline = new Pipeline().
setStages(Array(labelIndexer, featureAssembler))
val myFeatureTransfer = convpipeline.fit(df)
Apparently it didn't work. I am not sure what should I do to make the whole thing more automatic or ML pipeline does not take that many columns at this moment(which I doubt)?
I finally figured out one way, which is not very pretty. It is to create vector.dense for the features, and then create data frame out of this.
import org.apache.spark.mllib.regression.LabeledPoint
val myDataRDDLP = inputData.map {line =>
val indexed = line.split('\t').zipWithIndex
val myValues = indexed.filter(x=> {x._2 >1770}).map(x=>x._1).map(_.toDouble)
val mykey = indexed.filter(x=> {x._2 == 3}).map(x=>(x._1.toDouble-1)).mkString.toDouble
LabeledPoint(mykey, Vectors.dense(myValues))
}
val training = sqlContext.createDataFrame(myDataRDDLP).toDF("label", "features")
You shouldn't use quotes (s"$quote$x$quote") unless column names contain quotes. Try
val featureAssembler = new VectorAssembler().
setInputCols(featureArray).
setOutputCol("features")
For pyspark, you can first create a list of the column names:
df_colnames = df.columns
Then you can use that in vectorAssembler:
assemble = VectorAssembler(inputCols = df_colnames, outputCol = 'features')
df_vectorized = assemble.transform(df)
Related
I want to apply k-means for my parquet file.but error appear .
edited
java.lang.ArrayIndexOutOfBoundsException: 2
code
val Data = sqlContext.read.parquet("/usr/local/spark/dataset/norm")
val parsedData = Data.rdd.map(s => Vectors.dense(s.getDouble(1),s.getDouble(2))).cache()
import org.apache.spark.mllib.clustering.KMeans
val numClusters = 30
val numIteration = 1
val userClusterModel = KMeans.train(parsedData, numClusters, numIteration)
val userfeature1 = parsedData.first
val userCost = userClusterModel.computeCost(parsedData)
println("WSSSE for users: " + userCost)
How to solve this error?
I believe you are using https://spark.apache.org/docs/latest/mllib-clustering.html#k-means as a reference to build your K-Means model.
In the example
val data = sc.textFile("data/mllib/kmeans_data.txt")
val parsedData = data.map(s => Vectors.dense(s.split(' ').map(_.toDouble))).cache()
data is of type org.apache.spark.rdd.RDD In your case sqlContext.read.parquet is of type DataFrame. Therefore you would have to convert the dataframe to RDD to perform the split operation
To convert from Dataframe to RDD you can use the below sample as reference
val rows: RDD[Row] = df.rdd
val parsedData = Data.rdd.map(s => Vectors.dense(s.getInt(0),s.getDouble(1))).cache()
I'm trying to use Spark to learn multiclass logistic regression on a windowed text file. What I'm doing is first creating windows and explode them into $"word_winds". Then move the center word of each window into $"word". To fit the LogisticRegression model, I convert each different word into a class ($"label"), thereby it learns. I count the different labels to prone those with few minF samples.
The problem is that some part of the code is very very slow, even for small input files (you can use some README file to test the code). Googling, some users have been experiencing slowness by using explode. They suggest some modifications to the code in order to speed up 2x. However, I think that with a 100MB input file, this wouldn't be sufficient. Please suggest something different, probably to avoid actions that slow down the code. I'm using Spark 2.4.0 and sbt 1.2.8 on a 24-core machine.
import org.apache.spark.sql.functions._
import org.apache.spark.ml.classification.LogisticRegression
import org.apache.spark.ml.feature.{HashingTF, IDF}
import org.apache.spark.ml.feature.StringIndexer
import org.apache.spark.sql.SparkSession
import org.apache.spark.storage.StorageLevel
import org.apache.spark.sql.types._
object SimpleApp {
def main(args: Array[String]) {
val spark = SparkSession.builder().getOrCreate()
import spark.implicits._
spark.sparkContext.setCheckpointDir("checked_dfs")
val in_file = "sample.txt"
val stratified = true
val wsize = 7
val ngram = 3
val minF = 2
val windUdf = udf{s: String => s.sliding(ngram).toList.sliding(wsize).toList}
val get_mid = udf{s: Seq[String] => s(s.size/2)}
val rm_punct = udf{s: String => s.replaceAll("""([\p{Punct}|¿|\?|¡|!]|\p{C}|\b\p{IsLetter}{1,2}\b)\s*""", "")}
// Read and remove punctuation
var df = spark.read.text(in_file)
.withColumn("value", rm_punct($"value"))
// Creating windows and explode them, and get the center word into $"word"
df = df.withColumn("char_nGrams", windUdf('value))
.withColumn("word_winds", explode($"char_nGrams"))
.withColumn("word", get_mid('word_winds))
val indexer = new StringIndexer().setInputCol("word")
.setOutputCol("label")
df = indexer.fit(df).transform(df)
val hashingTF = new HashingTF().setInputCol("word_winds")
.setOutputCol("freqFeatures")
df = hashingTF.transform(df)
val idf = new IDF().setInputCol("freqFeatures")
.setOutputCol("features")
df = idf.fit(df).transform(df)
// Remove word whose freq is less than minF
var counts = df.groupBy("label").count
.filter(col("count") > minF)
.orderBy(desc("count"))
.withColumn("id", monotonically_increasing_id())
var filtro = df.groupBy("label").count.filter(col("count") <= minF)
df = df.join(filtro, Seq("label"), "leftanti")
var dfs = if(stratified){
// Create stratified sample 'dfs'
var revs = counts.orderBy(asc("count")).select("count")
.withColumn("id", monotonically_increasing_id())
revs = revs.withColumnRenamed("count", "ascc")
// Weigh the labels (linearly) inversely ("ascc") proportional NORMALIZED weights to word ferquency
counts = counts.join(revs, Seq("id"), "inner").withColumn("weight", col("ascc")/df.count)
val minn = counts.select("weight").agg(min("weight")).first.getDouble(0) - 0.01
val maxx = counts.select("weight").agg(max("weight")).first.getDouble(0) - 0.01
counts = counts.withColumn("weight_n", (col("weight") - minn) / (maxx - minn))
counts = counts.withColumn("weight_n", when(col("weight_n") > 1.0, 1.0)
.otherwise(col("weight_n")))
var fractions = counts.select("label", "weight_n").rdd.map(x => (x(0), x(1)
.asInstanceOf[scala.Double])).collectAsMap.toMap
df.stat.sampleBy("label", fractions, 36L).select("features", "word_winds", "word", "label")
}else{ df }
dfs = dfs.checkpoint()
val lr = new LogisticRegression().setRegParam(0.01)
val Array(tr, ts) = dfs.randomSplit(Array(0.7, 0.3), seed = 12345)
val training = tr.select("word_winds", "features", "label", "word")
val test = ts.select("word_winds", "features", "label", "word")
val model = lr.fit(training)
def mapCode(m: scala.collection.Map[Any, String]) = udf( (s: Double) =>
m.getOrElse(s, "")
)
var labels = training.select("label", "word").distinct.rdd
.map(x => (x(0), x(1).asInstanceOf[String]))
.collectAsMap
var predictions = model.transform(test)
predictions = predictions.withColumn("pred_word", mapCode(labels)($"prediction"))
predictions.write.format("csv").save("spark_predictions")
spark.stop()
}
}
Since your data is somewhat small it might help if you use coalesce before explode. Sometimes it can be inefficient to have too many nodes especially if there is a lot of shuffling in your code.
Like you said, it does seem like a lot of people have issues with explode. I looked at the link you provided but no one mentioned trying flatMap instead of explode.
I am trying to build a model in Spark ML with Zeppelin.
I am new to this area and would like some help. I think i need to set the correct datatypes to the column and set the first column as the label. Any help would be greatly appreciated, thank you
val training = sc.textFile("hdfs:///ford/fordTrain.csv")
val header = training.first
val inferSchema = true
val df = training.toDF
val lr = new LogisticRegression()
.setMaxIter(10)
.setRegParam(0.3)
.setElasticNetParam(0.8)
val lrModel = lr.fit(df)
// Print the coefficients and intercept for multinomial logistic regression
println(s"Coefficients: \n${lrModel.coefficientMatrix}")
println(s"Intercepts: ${lrModel.interceptVector}")
A snippet of the csv file i am using is:
IsAlert,P1,P2,P3,P4,P5,P6,P7,P8,E1,E2
0,34.7406,9.84593,1400,42.8571,0.290601,572,104.895,0,0,0,
As you have mentioned, you are missing the features column. It is a vector containing all predictor variables. You have to create it using VectorAssembler.
IsAlert is the label and all others variables (p1,p2,...) are predictor variables, you can create features column (actually you can name it anything you want instead of features) by:
import org.apache.spark.ml.feature.VectorAssembler
import org.apache.spark.ml.linalg.Vectors
//creating features column
val assembler = new VectorAssembler()
.setInputCols(Array("P1","P2","P3","P4","P5","P6","P7","P8","E1","E2"))
.setOutputCol("features")
val lr = new LogisticRegression()
.setMaxIter(10)
.setRegParam(0.3)
.setElasticNetParam(0.8)
.setFeaturesCol("features") // setting features column
.setLabelCol("IsAlert") // setting label column
//creating pipeline
val pipeline = new Pipeline().setStages(Array(assembler,lr))
//fitting the model
val lrModel = pipeline.fit(df)
Refer: https://spark.apache.org/docs/latest/ml-features.html#vectorassembler.
I am building a training set using two text files representing documents and labels.
Documents.txt
hello world
hello mars
Labels.txt
0
1
I have read in these files and converted my document data to a tf-idf weighted term-document matrix which is represented as a RDD[Vector]. I have also read-in and created a RDD[Vector] for my labels:
val docs: RDD[Seq[String]] = sc.textFile("Documents.txt").map(_.split(" ").toSeq)
val labs: RDD[Vector] = sc.textFile("Labels.txt")
.map(s => Vectors.dense(s.split(',').map(_.toDouble)))
val hashingTF = new HashingTF()
val tf: RDD[Vector] = hashingTF.transform(docs)
tf.cache()
val idf = new IDF(minDocFreq = 3).fit(tf)
val tfidf: RDD[Vector] = idf.transform(tf)
I would like to use tfidf and labsto create a RDD[LabeledPoint], but I am not sure how to apply a mapping with two different RDDs. Is this even possible/efficient, or do I need to rethink my approach?
One way to handle this is to join based on indices:
import org.apache.spark.RangePartitioner
// Add indices
val idfIndexed = idf.zipWithIndex.map(_.swap)
val labelsIndexed = labels.zipWithIndex.map(_.swap)
// Create range partitioner on larger RDD
val partitioner = new RangePartitioner(idfIndexed.partitions.size, idfIndexed)
// Join with custom partitioner
labelsIndexed.join(idfIndexed, partitioner).values
Below is the code I have for a RandomForest multiclass-classification model. I am reading from a CSV file and doing various transformations as seen in the code.
I am calculating the max number of categories and then giving it as a parameter to RF. This takes a lot of time! Is there a parameter to set or an easier way to make the model automatically infer the max categories?Since it can go more than 1000 and I cannot omit them.
How do I handle unseen labels on new data for prediction since StringIndexer will not work in that case. the code below is just a split of data but I will be introducing new data as well in future
// Need to predict 2 classes
val cols_to_predict=Array("Label1","Label2")
// ID col
val omit_cols=Array("Key")
// reading the csv file
val data = sqlContext.read
.format("com.databricks.spark.csv")
.option("header", "true") // Use first line of all files as header
.option("inferSchema", "true") // Automatically infer data types
.load("abc.csv")
.cache()
// creating a features DF by droppping the labels so that I can run all
// the cols through String Indexer
val features=data.drop("Label1").drop("Label2").drop("Key")
// Since I do not know my max categories possible, I find it out
// and use it for maxBins parameter in RF
val distinct_col_counts=features.columns.map(x => data.select(x).distinct().count ).max
val transformers: Array[org.apache.spark.ml.PipelineStage] = features.columns.map(
cname => new StringIndexer().setInputCol(cname).setOutputCol(s"${cname}_index").fit(features)
)
val assembler = new VectorAssembler()
.setInputCols(features.columns.map(cname => s"${cname}_index"))
.setOutputCol("features")
val labelIndexer2 = new StringIndexer()
.setInputCol("prog_label2")
.setOutputCol("Label2")
.fit(data)
val labelIndexer1 = new StringIndexer()
.setInputCol("orig_label1")
.setOutputCol("Label1")
.fit(data)
val rf = new RandomForestClassifier()
.setLabelCol("Label1")
.setFeaturesCol("features")
.setNumTrees(100)
.setMaxBins(distinct_col_counts.toInt)
val labelConverter = new IndexToString()
.setInputCol("prediction")
.setOutputCol("predictedLabel")
.setLabels(labelIndexer1.labels)
// Split into train and test
val Array(trainingData, testData) = data.randomSplit(Array(0.7, 0.3))
trainingData.cache()
testData.cache()
// Running only for one label for now Label1
val stages: Array[org.apache.spark.ml.PipelineStage] =transformers :+ labelIndexer1 :+ assembler :+ rf :+ labelConverter //:+ labelIndexer2
val pipeline=new Pipeline().setStages(stages)
val model=pipeline.fit(trainingData)
val predictions = model.transform(testData)