第4章 朴素贝叶斯算法实现中的先验概率问题

[junliang230]

您好，我对贝叶斯算法实现中的先验概率有些疑惑 `

计算概率

def calculate_probabilities(self, input_data):
    # summaries:{0.0: [(5.0, 0.37),(3.42, 0.40)], 1.0: [(5.8, 0.449),(2.7, 0.27)]}
    # input_data:[1.1, 2.2]
    probabilities = {}
    for label, value in self.model.items():
        probabilities[label] = 1 #probability[label]=1???
        for i in range(len(value)):
            mean, stdev = value[i]
            probabilities[label] *= self.gaussian_probability(input_data[i], mean, stdev) 
    return probabilities

`

为什么这里的probabilities[label]可以直接赋值为1呢，这样所有的类的先验概率是不是都一样了，为什么不根据样本计算这里的probabilities[label]呢 谢谢解答

[Xiaoccer]

同问

[k-burner]

同问……划分数据后类别不平衡了，先验概率应该不相等了吧

[hanhao0125]

是有问题，没有乘先验概率，作者不出来回应一下吗

[1344618323]

同问，希望能更改下

[josephcui]

确实少了先验概率，训练集的两类样本是不均衡的，一个是39，一个是31

[jiaozihao18]

同样有疑问，看了好久代码没找到先验概率计算。

[kpsc]

看了他给的链接，原始文章里面是写了先验概率的

[nicholaslsq]

This is Nicholas. I've received your mail.

[cyy0214]

Thank you. Your email is received and will be handled as soon as possible.Best Regards.This is an automatic reply,confirming that your e-mail was received.Thank you.

[xmingpyccnu]

您好！很开心收到您的邮件。我将在查看邮件后，尽快给您回复。

[daibitao19]

您好，我已经收到您的邮件，稍后我将尽快给你回复。