bayes

贝叶斯：

#!/usr/bin/python
# -*-coding:utf-8-*-
'''
Created on  Mar. 28, 2017
'''
import numpy as np

def getProb(data,label):
    if len(data)!=len(label):
        return False
    length=len(data)
    label1Count=0
    label0Count=0
    pro0Dict={}
    pro1Dict={}
    for i in range(length):
        if label[i]==-1:
            label0Count += 1
        if label[i]==1:
            label1Count += 1
    for i in range(length):
        if label[i] == -1:
            for j in range(len(data[0])):
                if j not in pro0Dict:
                    pro0Dict[j]={}
                if data[i][j] not in pro0Dict[j]:
                    pro0Dict[j][data[i][j]]=0.0
                pro0Dict[j][data[i][j]] += 1.0
        if label[i] == 1:
            for j in range(len(data[0])):
                if j not in pro1Dict:
                    pro1Dict[j] = {}
                if data[i][j] not in pro1Dict[j]:
                    pro1Dict[j][data[i][j]] = 0.0
                pro1Dict[j][data[i][j]] += 1.0
    for j in range(len(data[0])):
        for k in pro0Dict[j].keys():
            pro0Dict[j][k]=(pro0Dict[j][k]+1)/(label0Count+ len(pro0Dict[j].keys()))
        for k in pro1Dict[j].keys():
            pro1Dict[j][k] = (pro1Dict[j][k] + 1) / (label1Count + len(pro1Dict[j].keys()))
    pro1=float(label1Count+1)/(length+2)
    pro0=float(label0Count+1)/(length+2)
    return pro0, pro1,pro0Dict, pro1Dict

def predict(trainData,trainLabel,testData):
    pro0, pro1, pro0Dict, pro1Dict=getProb(trainData,trainLabel)
    p0= pro0* pro0Dict[0][str(testData[0])]*pro0Dict[1][str(testData[1])]
    p1 = pro1 * pro1Dict[0][str(testData[0])] * pro1Dict[1][str(testData[1])]
    print p0,p1
    if p0 >= p1:
        return -1
    else:
        return 1


if __name__ == "__main__":
    data=np.array([[1,'S'],[1,'M'],[1,'M'],[1,'S'],[1,'S'],[2,'S'],[2,'M'],[2,'M'],[2,'L'],[2,'L'],[3,'L'],[3,'M'],[3,'M'],[3,'L'],[3,'L']])
    label=np.array([-1,-1,1,1,-1,-1,-1,1,1,1,1,1,1,1,-1])
    test=[2,'S']
    print test[0]
    print getProb(data,label)
    print predict(data,label,test)