According to PCI(Programming Collective Intelligence),I've learned a lot.

def sim_pearson(prefs,p1,p2):

si={} for item in prefs[p1]: if item in prefs[p2]: si[item]=1#item是对应cristcs里面的电影名称，例如Snakes on a Plane

if len(si)==0: return 0#如果两者没有共同之处返回0

n=len(si)

sum1=sum([prefs[p1][it] for it in si]) sum2=sum([prefs[p2][it] for it in si])

sum1Sq=sum([pow(prefs[p1][it],2) for it in si]) sum2Sq=sum([pow(prefs[p2][it],2) for it in si])

pSum=sum([prefs[p1][it]*prefs[p2][it] for it in si])

num=pSum-(sum1sum2/n) den=sqrt((sum1Sq-pow(sum1,2)/n)(sum2Sq-pow(sum2,2)/n)) if den==0: return 0

r=num/den

return r #tp:根据皮尔逊公式等价关系重新编制 def sim_pearson_mean(prefs,p1,p2):

print("这是皮尔逊另外一个等价公式计算函数") si={} for item in prefs[p1]: if item in prefs[p2]: si[item]=1#item是对应cristcs里面的电影名称，例如Snakes on a Plane

if len(si)==0: return 0#如果两者没有共同之处返回0

n=len(si)

#计算平均值 mean1=sum([prefs[p1][it] for it in si])/n mean2=sum([prefs[p2][it] for it in si])/n

#计算协方差 cov=sum([(prefs[p1][it]-mean1)*(prefs[p2][it]-mean2) for it in si])

#计算平方和 sum1s=sum([pow((prefs[p1][it]-mean1),2) for it in si]) sum2s=sum([pow((prefs[p2][it]-mean2),2) for it in si])

#计算分母 sumM=sqrt(sum1s*sum2s)

if sumM==0: return 0

r=cov/sumM

return r