Python提取高频词作为话题词_python词汇话题聚类

import pandas as pd
from collections import Counter
from sklearn.feature_extraction.text import TfidfVectorizer
import spacy

# 加载spaCy的英语模型
nlp = spacy.load("en_core_web_sm")
nlp.max_length = 1500000  # 设置一个更大的值，根据你的文本长度调整

# 读取Excel文件
file_path = 'cleaned_Laos_news.xlsx'
sheet_name = 'Sheet1'  # 修改为实际的工作表名称
data = pd.read_excel(file_path, sheet_name=sheet_name)

# 清除NaN值，将它们替换为一个空字符串
data['cleaned_content'].fillna('', inplace=True)

# 合并所有新闻文本内容
all_texts = data['cleaned_content'].tolist()

# 使用spaCy进行分词和关键词提取
docs = list(nlp.pipe(all_texts))

# 统计高频词
topic_keywords = [token.text for doc in docs for token in doc if not token.is_stop and len(token.text) > 1 and not token.is_punct]
word_counts = Counter(topic_keywords)

# 计算TF-IDF权重
tfidf_vectorizer = TfidfVectorizer()
tfidf_matrix = tfidf_vectorizer.fit_transform(all_texts)

# 获取词汇表和对应的词语
vocabulary = tfidf_vectorizer.get_feature_names_out()

# 定义权重
tfidf_weight = 0.7  # TF-IDF权重
word_count_weight = 0.3  # 高频词得分权重

# 综合得分并选择话题词
topic_scores = {}
for i, word in enumerate(vocabulary):
    tfidf_score = tfidf_matrix[:, i].sum()
    word_score = word_counts[word]
    combined_score = tfidf_weight * tfidf_score + word_count_weight * word_score
    topic_scores[word] = combined_score

# 选择综合得分最高的词作为话题词
top_n = 20
topic_keywords = sorted(topic_scores.keys(), key=lambda x: topic_scores[x], reverse=True)[:top_n]

print("提取的综合得分高频关键词作为话题词：")
print(topic_keywords)