机器学习复现3.朴素贝叶斯法_机器学习蛋白组文章复现-CSDN博客

本文链接：https://blog.csdn.net/qq_50438796/article/details/128751271

文章展示了一个使用Python编写的朴素贝叶斯分类器示例。该模型基于给定的训练数据集，利用Counter处理数据并进行预测。分类器考虑了正则化项（lamda），可以进行极大似然估计或贝叶斯估计。在给定测试数据后，模型能预测其类别。

摘要生成于 C知道，由 DeepSeek-R1 满血版支持，前往体验 >

import numpy as np
from collections import Counter


class Naive_Bayes(object):
    dataset = [{'x':[1,'S'],'y':-1},{'x':[1,'M'],'y':-1},
        {'x':[1,'M'],'y':1},{'x':[1,'S'],'y':1},
        {'x':[1,'S'],'y':-1},{'x':[2,'S'],'y':-1},
        {'x':[2,'M'],'y':-1},{'x':[2,'M'],'y':1},
        {'x':[2,'L'],'y':1},{'x':[2,'L'],'y':1},
        {'x':[3,'L'],'y':1},{'x':[3,'M'],'y':1},
        {'x':[3,'M'],'y':1},{'x':[3,'L'],'y':1},
        {'x':[3,'L'],'y':-1}]
    dim = 0   ## 维度
    lamda = 0    ## 正则化项，lamda = 0，相当于极大似然估计，lamda大于0，相当于贝叶斯估计，先验概率有效。
    prior_posibility = []
    condition_posibility = []

    def __init__(self):
        total_len = len(self.dataset)
        self.dim = len(self.dataset[0]['x'])
        counter_y = Counter(i['y'] for i in self.dataset)
        for i in counter_y.keys():
            self.prior_posibility.append(dict(mkey = i,value = (counter_y[i] + self.lamda)/(total_len + len(counter_y) * self.lamda)))
        tmp = []
        for i in range(self.dim):  ## 维度
            tmp1 = []
            counter_total_x = Counter(w['x'][i] for w in self.dataset)
            for s in counter_y.keys():
                tmp2 = []
                counter_x = Counter(w['x'][i] for w in self.dataset if w['y'] == s)
                for p in counter_total_x.keys():
                    tmp2.append(dict(mkey=p, value=(counter_x[p] + self.lamda) / (counter_y[s] + len(counter_total_x) * self.lamda)))

                tmp1.append(dict(mkey = s,value = tmp2))
            tmp.append(tmp1)
        self.condition_posibility = tmp

    def predict(self,test_data):
        posibility = []
        for result in range(len(self.prior_posibility)):
            pos = self.prior_posibility[result]['value']
            for i in range(k.dim):
                for w in self.condition_posibility[i][result]['value']:
                    if(w['mkey'] == test_data[i]):
                        pos *= w['value']
                        break
            posibility.append(pos)
        return self.prior_posibility[np.argmax(posibility)]['mkey']


k = Naive_Bayes()
posibility = []
test = [2,'S']
print(test,"的预测结果为: ",k.predict(test))