模式识别小课设：基于 TF-IDF 与 Logistic 回归的新闻文本五分类引擎

代码如下：

import os
import numpy as np
import pandas as pd
from sklearn.feature_extraction.text import TfidfVectorizer
from sklearn.linear_model import LogisticRegression
from sklearn.metrics import accuracy_score

# 定义新闻类别
categories = ["军事", "经济", "文化", "科学", "娱乐"]

# 文件路径
data_folder = "E:/news"

# 读取数据函数
def load_data(data_folder, categories, train_range, test_range):
    train_texts, train_labels = [], []
    test_texts, test_labels = [], []
    for category in categories:
        folder_path = os.path.join(data_folder, category)
        for i in range(1, 26):  # 文件从1到25
            file_path = os.path.join(folder_path, f"{i}.txt")
            with open(file_path, 'r', encoding='utf-8') as file:
                content = file.read()
                if i in train_range:
                    train_texts.append(content)
                    train_labels.append(category)
                elif i in test_range:
                    test_texts.append(content)
                    test_labels.append(category)
    return train_texts, train_labels, test_texts, test_labels

# 加载数据
train_range = range(1, 21)  # 训练集范围：1-20
test_range = range(21, 26)  # 测试集范围：21-25
train_texts, train_labels, test_texts, test_labels = load_data(data_folder, categories, train_range, test_range)

# 特征提取
vectorizer = TfidfVectorizer(max_features=5000)  # 限制最大特征数
X_train = vectorizer.fit_transform(train_texts)
X_test = vectorizer.transform(test_texts)

# 标签映射
label_mapping = {category: i for i, category in enumerate(categories)}
y_train = np.array([label_mapping[label] for label in train_labels])
y_test = np.array([label_mapping[label] for label in test_labels])

# 训练逻辑回归模型
model = LogisticRegression(max_iter=1000, solver='lbfgs', multi_class='ovr')
model.fit(X_train, y_train)

# 测试预测
y_pred = model.predict(X_test)

# 计算准确率
accuracy = accuracy_score(y_test, y_pred)
print("Logistic回归模型测试集准确率:", accuracy)
print(y_pred)
print(y_test)

结果如下：

正确率达到80%，初步估计是因为数据太少，这里只选用了120份数据。采用更多数据学习效果会更好。