贝叶斯分类器（续近邻分类器）

#A Naive Bayesian Classifier


total ={} #类训练实例,存储对应类的出现的次数
histo ={} #存储对应类中的对应特征的值得频率


train = open("D:\\iris.trn",'r')


for line in train:
    f = line.rstrip().split(',')
    
    label = f.pop()
    
    if not total.has_key(label):
        total[label] =0
        histo[label]=[{},{},{},{}]
    
    total[label] +=1
    
    for i in range(4):
        histo[label][i][f[i]] = 1 +histo[label][i].get(f[i],0.0)
train.close()
        
#读取测试集并且评估可能性，选出最大可能性的类
hit , miss = 0,0
test = open("D:\\iris.tst")


for line in test:
    f = line.rstrip().split(',')
    true = f.pop()
    
    p = {} #类的可能性
    
    for label in total.keys():
        p[label] =1 
        for i in range(4):
            p[label] *= histo[label][i].get(f[i],0.0)/total[label] #计算出类中的对应的属性的频率的乘积


    mx ,predicted = 0,-1;
    for k in p.keys():#找出最大的概率
        if p[k] >=mx:
            mx,predicted=p[k],k
    
    if true == predicted:
        flag ='+'
        hit +=1
    else:
        flag ='-'
        miss +=1
    print flag ,"\t",true,"\t",predicted,"\t",
    
    for label in p.keys():
        print label,":",p[label],"\t",
    print
print
print hit,"out of ",hit+miss,"correct-Accuracy: ",hit/(hit+miss+0.0)
test.close()

结果

+ Iris-setosa Iris-setosa Iris-virginica : 0.0 Iris-setosa : 0.000764069733796 Iris-versicolor : 0.0
+ Iris-setosa Iris-setosa Iris-virginica : 0.0 Iris-setosa : 0.000377136983989 Iris-versicolor : 0.0
+ Iris-versicolor Iris-versicolor Iris-virginica : 0.0 Iris-setosa : 0.0 Iris-versicolor : 2.88e-05
+ Iris-versicolor Iris-versicolor Iris-virginica : 0.0 Iris-setosa : 0.0 Iris-versicolor : 0.0004368
+ Iris-virginica Iris-virginica Iris-virginica : 1.728e-05 Iris-setosa : 0.0 Iris-versicolor : 0.0


5 out of  5 correct-Accuracy:  1.0