朴素贝叶斯算法实现python

朴素贝叶斯算法是一种基于概率统计的分类算法，常用于文本分类、垃圾邮件过滤等任务。下面是一个简单的朴素贝叶斯算法的Python实现示例：

import numpy as np

class NaiveBayes:
    def __init__(self):
        self.prior_prob = {}  # 存储先验概率
        self.cond_prob = {}  # 存储条件概率

    def train(self, X, y):
        # 计算先验概率
        classes, counts = np.unique(y, return_counts=True)
        total_samples = len(y)
        for i in range(len(classes)):
            self.prior_prob[classes[i]] = counts[i] / total_samples

        # 计算条件概率
        for c in classes:
            self.cond_prob[c] = {}
            c_indices = np.where(y == c)[0]
            c_X = X[c_indices]
            total_words = np.sum(c_X)
            for j in range(X.shape[1]):
                word_count = np.sum(c_X[:, j])
                self.cond_prob[c][j] = (word_count + 1) / (total_words + X.shape[1])

    def predict(self, X):
        y_pred = []
        for x in X:
            max_prob = -1
            pred_class = None
            for c in self.prior_prob.keys():
                prob = np.log(self.prior_prob[c])
                for i in range(len(x)):
                    if x[i] > 0:
                        prob += np.log(self.cond_prob[c][i])
                if prob > max_prob:
                    max_prob = prob
                    pred_class = c
            y_pred.append(pred_class)
        return y_pred

上述代码实现了一个简单的朴素贝叶斯分类器，包括训练和预测两个方法。在训练阶段，先计算各个类别的先验概率，然后计算每个词对于每个类别的条件概率。在预测阶段，根据先验概率和条件概率计算样本属于每个类别的概率，并选择概率最大的类别作为预测结果。