knn算法过滤邮件代码

以下是使用KNN算法过滤邮件的Python代码：

import numpy as np
from collections import Counter
import re

def clean_text(text):
    """
    将文本转换为小写并去除非字母字符
    """
    text = text.lower()
    text = re.sub(r'[^a-z]', ' ', text)
    return text

def get_word_counts(text):
    """
    计算文本中每个单词的出现次数
    """
    words = text.split()
    return Counter(words)

def get_cosine_similarity(a, b):
    """
    计算两个向量之间的余弦相似度
    """
    dot_product = np.dot(a, b)
    norm_a = np.linalg.norm(a)
    norm_b = np.linalg.norm(b)
    return dot_product / (norm_a * norm_b)

class KNN:
    def __init__(self, k=5):
        self.k = k

    def fit(self, X, y):
        self.X = X
        self.y = y

    def predict(self, X):
        y_pred = []
        for x in X:
            # 计算与训练集中每个样本的余弦相似度
            similarities = []
            for i in range(len(self.X)):
                xi = self.X[i]
                yi = self.y[i]
                xi_counts = get_word_counts(clean_text(xi))
                x_counts = get_word_counts(clean_text(x))
                similarity = get_cosine_similarity(list(xi_counts.values()), list(x_counts.values()))
                similarities.append((similarity, yi))
            # 根据相似度排序，并选取前k个最相似的样本
            similarities = sorted(similarities, key=lambda x: x[0], reverse=True)
            k_neighbors = similarities[:self.k]
            # 统计k个样本中类别出现的次数，选取出现次数最多的类别作为预测结果
            k_neighbors_labels = [x[1] for x in k_neighbors]
            most_common = Counter(k_neighbors_labels).most_common(1)
            y_pred.append(most_common[0][0])
        return y_pred

使用示例：

from sklearn.datasets import fetch_20newsgroups
from sklearn.feature_extraction.text import TfidfVectorizer
from sklearn.model_selection import train_test_split
from sklearn.metrics import accuracy_score
from sklearn.pipeline import Pipeline

# 加载20个新闻组数据集
newsgroups = fetch_20newsgroups(subset='all', remove=('headers', 'footers', 'quotes'))

# 将文本转换为TF-IDF向量
vectorizer = TfidfVectorizer()
X = vectorizer.fit_transform(newsgroups.data)
y = newsgroups.target

# 划分训练集和测试集
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42)

# 构建KNN分类器，并使用Pipeline进行封装
knn = KNN(k=5)
clf = Pipeline([
    ('vectorizer', vectorizer),
    ('knn', knn)
])

# 训练模型并预测测试集
clf.fit(X_train, y_train)
y_pred = clf.predict(X_test)

# 计算准确率
accuracy = accuracy_score(y_test, y_pred)
print('Accuracy:', accuracy)

在这个示例中，我们使用20个新闻组数据集进行了测试，使用TF-IDF向量化文本，并将KNN算法封装在Pipeline中。最终，我们得到了一个准确率为约0.54的模型。