Getting Error using 'Linearregression' in spark 2.3.4

#Load the csv data into a RDD autoData= sc.textFile("auto-miles-per-gallon.csv") autoData.cache() #Remove the first line that contains header dataLines = autoData.filter(lambda x: "CYLINDERS" not in x) dataLines.count() # Convert the RDD into a Dense Vector. As a part of this excercise we'll # 1. Remove unwanted columns # 2. change non-numeric values to numeric import math from pyspark.mllib.linalg import Vectors #Use default for average HP avgHP=sc.broadcast(80.0) def transformToNumeric(inputStr):     global avgHP     attList= inputStr.split(",")          #Replace ? values with a normal value     hpvalue= attList[3]     if hpvalue =="?":         hpvalue= avgHP.value     # Filterout the columns which are not needed at this stage     values = Vectors.dense([float(attList[0]), float(attList[1]), \                             hpvalue, \                             float(attList[5]),                             float(attList[6])])     return values #Keep only MPG, CYLINDERS, HP, ACCELERATION and MODEL YEAR autoVectors = dataLines.map(transformToNumeric) autoVectors.collect() # Perform statistical analysis from pyspark.mllib.stat import Statistics autoStats = Statistics.colStats(autoVectors) autoStats.mean() autoStats.variance() autoStats.min() autoStats.max() Statistics.corr(autoVectors) #Transform to a dtatframe for input to machine learning # Drop columns that are nor required (Low-correlation) from pyspark.sql import SQLContext sqlContext = SQLContext(sc) def transformToLabeledPoint(inStr):     # We're ignoring 3rd column because of low co-relation value     lp= ( float(inStr[0]), Vectors.dense([inStr[1],inStr[2],inStr[4]]))     return lp autoLp = autoVectors.map (transformToLabeledPoint) #this will be a new RDD autoDF = sqlContext.createDataFrame(autoLp,["label","features"]) autoDF.select("label","features").show(10) #Find Corrrelations numFeatures = autoDF.take(1)[0].features.size labelRDD =autoDF.rdd.map(lambda lp: float(lp.label)) for i in range(numFeatures):     featureRDD = autoDF.rdd.map(lambda lp: lp.features[i])     corr = Statistics.corr(labelRDD,featureRDD, 'pearson')     print("%d\t%g" % (i,corr))      # Split into Traning and Testing data (trainingData,testData) =autoDF.randomSplit([0.9,0.1]) trainingData.count() testData.count() # Build the Model on training data from pyspark.ml.regression import LinearRegression lr = LinearRegression(maxIter =10) lrModel = lr.fit(trainingData)