Distributed implementation of logistic regression [logistic regression/machine Learning/spark]

1-Questions raised

2-Logistic regression

3-Theoretical derivation

4-python/spark implementation

1 #-*-coding:utf-8-*-2  fromPysparkImportSparkcontext3  fromMathImport*4 5theta = [0, 0, 0]#Initial theta Value6Alpha = 0.001#Learning Rate7 8 definner (x, y):9     returnSUM ([i*j forI,jinchzip (x, y)])Ten          One deffunc (LST): AH = (1 + exp (-inner (LST, theta))) * * (-1) -     returnMapLambdax: (H-lst[-1]) * x, Lst[:-1]) -  the  -sc = Sparkcontext ('Local') -  -Rdd = Sc.textfile ('/home/freyr/logisticregression.txt') +. Map (LambdaLine:map (float, Line.strip (). Split (','))) -. Map (LambdaLST: [1]+lst) +  A  at  forIinchRange (400): -Partheta =Rdd.map (func) -. Reduce (LambdaX, y: [i+j forI,jinchzip (x, y)]) -  -      forJinchRange (3): -THETA[J] = theta[j]-Alpha *Partheta[j] in  - Print 'theta =%s'% Theta

PS:logisticRegression.txt

Distributed implementation of logistic regression [logistic regression/machine Learning/spark]