朴素贝叶斯python代码实例垃圾邮件

当然，以下是一个基于Python的朴素贝叶斯垃圾邮件分类器的示例代码：

import os
import io
import numpy as np
from collections import Counter

# 创建一个函数用于读取文件内容并返回一个单词列表
def get_words_in_file(file):
    with io.open(file, 'r', encoding='utf-8') as f:
        return f.read().split()

# 创建一个函数用于读取文件夹中所有的文件
def get_files_in_folder(folder):
    files = []
    for (dirpath, dirnames, filenames) in os.walk(folder):
        for filename in filenames:
            files.append(os.path.join(dirpath, filename))
    return files

# 创建一个函数用于将单词列表转换为一个字典
def get_word_counts(words):
    return dict(Counter(words))

# 创建一个函数用于计算每个类别（垃圾邮件和非垃圾邮件）中每个单词出现的次数
def calculate_class_word_counts(training_files):
    class_word_counts = {}
    for file in training_files:
        if 'ham' in file:
            if 'ham' not in class_word_counts:
                class_word_counts['ham'] = {}
            words = get_words_in_file(file)
            word_counts = get_word_counts(words)
            for word, count in word_counts.items():
                if word not in class_word_counts['ham']:
                    class_word_counts['ham'][word] = 0
                class_word_counts['ham'][word] += count
        else:
            if 'spam' not in class_word_counts:
                class_word_counts['spam'] = {}
            words = get_words_in_file(file)
            word_counts = get_word_counts(words)
            for word, count in word_counts.items():
                if word not in class_word_counts['spam']:
                    class_word_counts['spam'][word] = 0
                class_word_counts['spam'][word] += count
    return class_word_counts

# 创建一个函数用于计算每个类别（垃圾邮件和非垃圾邮件）中所有单词的总数
def calculate_class_total_words(class_word_counts):
    total_words = {'ham': 0, 'spam': 0}
    for class_name, words in class_word_counts.items():
        for word, count in words.items():
            total_words[class_name] += count
    return total_words

# 创建一个函数用于计算每个类别（垃圾邮件和非垃圾邮件）中每个单词的概率
def calculate_class_word_probabilities(class_word_counts, total_words):
    class_word_probabilities = {}
    for class_name, words in class_word_counts.items():
        class_word_probabilities[class_name] = {}
        for word, count in words.items():
            class_word_probabilities[class_name][word] = count / total_words[class_name]
    return class_word_probabilities

# 创建一个函数用于预测新邮件的类别
def predict_new_email_class(new_email_file, class_word_probabilities, total_words):
    words = get_words_in_file(new_email_file)
    word_counts = get_word_counts(words)
    spam_probability = 1
    ham_probability = 1
    for word, count in word_counts.items():
        if word in class_word_probabilities['spam']:
            spam_probability *= (class_word_probabilities['spam'][word] ** count)
        if word in class_word_probabilities['ham']:
            ham_probability *= (class_word_probabilities['ham'][word] ** count)
    spam_probability *= total_words['spam'] / (total_words['spam'] + total_words['ham'])
    ham_probability *= total_words['ham'] / (total_words['spam'] + total_words['ham'])
    if spam_probability > ham_probability:
        return 'spam'
    else:
        return 'ham'

# 创建一个函数用于计算分类器的准确性
def calculate_classifier_accuracy(testing_files, class_word_probabilities, total_words):
    num_correct = 0
    for file in testing_files:
        predicted_class = predict_new_email_class(file, class_word_probabilities, total_words)
        if ('ham' in file and predicted_class == 'ham') or ('spam' in file and predicted_class == 'spam'):
            num_correct += 1
    return num_correct / len(testing_files)

# 读取训练数据和测试数据
training_files = get_files_in_folder('training')
testing_files = get_files_in_folder('testing')

# 计算每个类别（垃圾邮件和非垃圾邮件）中每个单词出现的次数
class_word_counts = calculate_class_word_counts(training_files)

# 计算每个类别（垃圾邮件和非垃圾邮件）中所有单词的总数
total_words = calculate_class_total_words(class_word_counts)

# 计算每个类别（垃圾邮件和非垃圾邮件）中每个单词的概率
class_word_probabilities = calculate_class_word_probabilities(class_word_counts, total_words)

# 计算分类器的准确性
accuracy = calculate_classifier_accuracy(testing_files, class_word_probabilities, total_words)

# 输出分类器的准确性
print('Accuracy:', accuracy)

在这个代码示例中，我们使用了Python的标准库来读取文件和处理文本数据。我们还使用了Python的Counter类来计算每个单词在文件中出现的次数，以及字典来存储每个类别中每个单词的出现次数和概率。最后，我们使用了准确性作为评估分类器的指标。