python实现贝叶斯

data = {'s':4,'f':7,'e':8}
print(max(data,key = data.get)) #输出为e

#读取数据
def loadData(filename):
    with open(filename) as fr:
        dataSet = fr.readlines()  #读取数据，每一行为列表中的元素
        dataSetMat = []
        labels = []
        for line in dataSet:
            line = line.strip().split('  ')  #去空格，分割
            labels.append(line[-1])
            dataSetMat.append(line)
        return dataSetMat,labels
dataSet,labels = loadData('data.txt')

def train(dataSet,labels):
    unique_labels = set(labels)
    res = {}
    #分别考虑每个类 书上151页 （机器学习 周志华）
    for label in unique_labels:
        res[label] = []
        res[label].append(labels.count(label)/float(len(labels))) #每个类的概率
        #每个属性下的个个取值的概率
        for i in range(len(dataSet[0]) - 1):
            tempCol = [l[i] for l in dataSet if l[-1] == label]  #对应类标签下的某个属性下的列取值
            uniqueVlaue = set(tempCol)
            dict = {}
            for value in uniqueVlaue:
                count = tempCol.count(value)
                prob = count/float(labels.count(label)) #计算的是某个类下属性取值为value的概率
                dict[value] = prob
            res[label].append(dict)
    return res
#{'soft': [0.3333333333333333, {'3': 0.375, '1': 0.25, '2': 0.375}, {'1': 0.375, '2': 0.625}, {'1': 0.875, '2': 0.125}, {'1': 0.25, '2': 0.75}, {'3': 0.5, '1': 0.125, '2': 0.375}], 'no': [0.4583333333333333, {'3': 0.36363636363636365, '1': 0.36363636363636365, '2': 0.2727272727272727}, {'1': 0.5454545454545454, '2': 0.45454545454545453}, {'1': 0.45454545454545453, '2': 0.5454545454545454}, {'1': 0.9090909090909091, '2': 0.09090909090909091}, {'3': 0.36363636363636365, '1': 0.36363636363636365, '2': 0.2727272727272727}], 'hard': [0.20833333333333334, {'3': 0.2, '1': 0.4, '2': 0.4}, {'1': 0.6, '2': 0.4}, {'2': 1.0}, {'2': 1.0}, {'1': 0.6, '2': 0.4}]}

#就是字典的索引
def test(testVect,probMat): #proMat可就是上边的那个字典
    hard = probMat['hard'] #每个类对应的后面的值
    soft = probMat['soft']
    no = probMat['no']
    #每个类的概率
    phard = hard[0]
    psoft = soft[0]
    pno = no[0]

    #循环测试数据
    for i in range(len(testVect)):
        if testVect[i] in hard[i+1].keys():
            phard *= hard[i+1][testVect[i]]
        else:
            phard = 0

        if testVect[i]  in soft[i+1].keys():
            psoft *= soft[i+1][testVect[i]]

        else:
            psoft = 0

        if testVect[i] in no[i+1].keys():
            pno *= no[i+1][testVect[i]]
        else:
            pno = 0
    res = {}
    res['hard'] = phard
    res['soft'] = psoft
    res['no'] = pno
    print(phard,psoft,pno)
    return max(res,key = res.get)

probMat = train(dataSet,labels)
res = test(['3','1','2','2','1'],probMat)
print(res)