Function map

src/python/15.BigData_MapReduce/proximalSVM.py:11–33

(key, value)

import numpy

def map(key, value):
   # input key= class for one training example, e.g. "-1.0"
   classes = [float(item) for item in key.split(",")]   # e.g. [-1.0]
   D = numpy.diag(classes)

   # input value = feature vector for one training example, e.g. "3.0, 7.0, 2.0"
   featurematrix = [float(item) for item in value.split(",")]
   A = numpy.matrix(featurematrix)

   # create matrix E and vector e
   e = numpy.matrix(numpy.ones(len(A)).reshape(len(A), 1))
   E = numpy.matrix(numpy.append(A, -e, axis=1)) 

   # create a tuple with the values to be used by reducer
   # and encode it with base64 to avoid potential trouble with '\t' and '\n' used
   # as default separators in Hadoop Streaming
   producedvalue = base64.b64encode(pickle.dumps( (E.T*E, E.T*D*e))    

   # note: a single constant key "producedkey" sends to only one reducer
   # somewhat "atypical" due to low degree of parallism on reducer side
   print "producedkey\t%s" % (producedvalue)

def reduce(key, values, mu=0.1):
  sumETE = None
  sumETDe = None