import pandas as pd

import re import jieba fp=open('dijiu.txt','r',encoding='UTF-8') data=fp.read() fp.close() data=data.replace(' ','') data=data.replace(',','') data=data.replace(' ','') data=data.replace('，','') data=data.replace('中国','') data=data.replace('全会','') data=data.split("。") #data = pd.Dataframe({'正文':data,}) data=pd.Dataframe(list(data),columns=['评论']) data_cut=data['评论'].apply(jieba.lcut) stop_word = pd.read_csv('stopword.txt',sep='hhe') stop_words = list(stop_word.iloc[:,0]) + [' '] data_after_stop = data_cut .apply(lambda x:[i for i in x if i not in stop_words]) index = data_after_stop.apply(lambda x:len(x)!=0) data_after_stop_not_null = data_after_stop[index] dic={} for i in data_after_stop: for j in i: if j not in dic.keys(): dic[j]=1 else: dic[j]+=1 score = pd.read_csv('BosonNLP_sentiment_score.txt',sep=' ',header=None) score.columns = ['word','word_score'] degree = pd.read_csv('degree.csv') degree.columns = ['degree_word','degree_score'] not_words = pd.read_csv('not.csv') not_words.columns = ['not_word'] not_words['not_word_score'] = -1 degree['degree_score'] = -degree['degree_score']/100 def get_score(x=None): t = pd.Dataframe(x) t.columns = ['word2'] t_score = pd.merge(t,score,how='left',left_on='word2',right_on='word') tmp = pd.merge(t_score,degree,how='left',left_on='word2',right_on='degree_word') ind = tmp['degree_word'].notnull() if a != (len(tmp)-1): tmp.loc[a+1,'word_score']=tmp.loc[a,'word_score']*tmp.loc[a+1,'degree_score'] tmp1 = pd.merge(tmp,not_words,how='left',left_on='word2',right_on='not_word') ind = tmp1['not_word'].notnull() if a!= (len(tmp1)-1): tmp1.loc[a+1,'word_score']=tmp1.loc[a+1,'word_score']*tmp1.loc[a,'not_word_score'] return tmp1['word_score'].sum() y=data_after_stop_not_null.apply(get_score) messge_ = data_after_stop_not_null.apply(lambda x:' '.join(x)) inde=y>0    以上就是本篇文章【LDA_LDA关键词_主题词提取_】的全部内容了，欢迎阅览 ！ 文章地址：http://mdekt.bhha.com.cn/news/98.html 
     资讯      企业新闻      行情      企业黄页      同类资讯      首页      网站地图      返回首页 康宝晨资讯移动站 http://weazh.bhha.com.cn/ , 查看更多