Classification

classification.py

'''Now given a mail, split it in terms of spaces  ,  then ,  add up the log probability of each .  Multiply it with the spam probability  . Do the same thing for non-spam 
   Whichever is higher  wins .  Lets start
'''
import sys,os
def makeDict(f):
	temp = {}
	data = open(f,'r')
	for line in data:
		prob = line.split(" ")
		temp[prob[0]] = prob[1]
	return temp
spamProbs = makeDict(sys.argv[1]) #Pass the spam log probs here 
nonspamProbs = makeDict(sys.argv[2]) #Pass the non-spam log probs here

toClassify = open(sys.argv[3],'r')
for line in toClassify:
	words = line.split(" ")
	
	spamP = 0
	nonspamP = 0
	for w in words:
		try:
			spamP = spamP + float(spamProbs[w].strip("\n"))
		except:
			
			spamP = spamP + 1
		try:
			nonspamP = nonspamP + float(nonspamProbs[w].strip("\n"))
		except:
			
			nonspamP = nonspamP + 1
	totalSpamP = spamP * 0.5
	totalnonSpamP = nonspamP * 0.5
	
	if(totalSpamP > totalnonSpamP):
		print 'This mail is spam'
	else:
		print 'This mail is not spam'