Python實現樸素貝葉斯分類器

# -*-coding:utf-8-*-
'''
    樸素貝葉斯算法
'''


from __future__ import division 


global className
className = "class"


def calc_class(train, classValue):
    # 計算分類的概率
    _num_cls = 0
    _num_trains = len(train)
    for t in train:
        if t[className] == classValue:
            _num_cls += 1
    return  _num_cls / _num_trains
    
def calc_attr(train, classValue, attrName, attrValue):
    # 計算屬性的概率
    _num_cls  = 0
    _num_attr = 0
    for a in train:
        if a[className] == classValue:
            _num_cls += 1
            if a[attrName] == attrValue:
                _num_attr += 1
                
    if _num_attr == 0:
        _num_attr = 1
    return _num_attr / _num_cls


def bayes(train, test, classY, classN):
    
    _prob_Y = calc_class(train, classY)
    _prob_N = calc_class(train, classN)
    for key,value in test.items():
        _prob_Y *= calc_attr(train, classY, key, value)
        _prob_N *= calc_attr(train, classN, key, value)
    
    return {classY:_prob_Y,classN:_prob_N}
    


if __name__=='__main__':
    # 訓練數據
    train = [  
             {"outlook":"sunny",    "temp":"hot",  "humidity":"high",   "wind":"weak",   "class":"no" },  
             {"outlook":"sunny",    "temp":"hot",  "humidity":"high",   "wind":"strong", "class":"no" },  
             {"outlook":"overcast", "temp":"hot",  "humidity":"high",   "wind":"weak",   "class":"yes" },  
             {"outlook":"rain",     "temp":"mild", "humidity":"high",   "wind":"weak",   "class":"yes" },  
             {"outlook":"rain",     "temp":"cool", "humidity":"normal", "wind":"weak",   "class":"yes" },  
             {"outlook":"rain",     "temp":"cool", "humidity":"normal", "wind":"strong", "class":"no" },  
             {"outlook":"overcast", "temp":"cool", "humidity":"normal", "wind":"strong", "class":"yes" },  
             {"outlook":"sunny",    "temp":"mild", "humidity":"high",   "wind":"weak",   "class":"no" },  
             {"outlook":"sunny",    "temp":"cool", "humidity":"normal", "wind":"weak",   "class":"yes" },  
             {"outlook":"rain",     "temp":"mild", "humidity":"normal", "wind":"weak",   "class":"yes" },  
             {"outlook":"sunny",    "temp":"mild", "humidity":"normal", "wind":"strong", "class":"yes" },  
             {"outlook":"overcast", "temp":"mild", "humidity":"high",   "wind":"strong", "class":"yes" },  
             {"outlook":"overcast", "temp":"hot",  "humidity":"normal", "wind":"weak",   "class":"yes" },  
             {"outlook":"rain",     "temp":"mild", "humidity":"high",   "wind":"strong", "class":"no" },  
             ] 
    # 測試數據 
    test = {"outlook":"overcast","temp":"cool","humidity":"high","wind":"strong"}
      
    print bayes(train, test, "yes", "no")