2つの文字列のcos類似度を測る

cosSim.py

# -*- coding:utf-8 -*-

import math

#cos類似度
def cosSim(s1,s2):
    v1 = dict([(word, s1.count(word)) for word in s1 + s2])   #s1の各キャラクタと頻度
    v2 = dict([(word, s2.count(word)) for word in s1 + s2])   #s2の各キャラクタと頻度

    print v1
    print v2

    nume = sum([i*j for i,j in zip(v1.values(), v2.values())])                              #分子
    deno = math.sqrt(sum([i**2 for i in v1.values()]) * sum([i**2 for i in v2.values()]))   #分母

    return nume / deno

s1 = 'a cat sat on the mat'
s2 = 'cats are sitting on the mat'

print cosSim(s1.split(' '), s2.split(' '))