基于LDA的网络舆情分析智能平台的设计与实现

Design and Implementation of an Intelligent Platform for Network Public Opinion Analysis Based on LDA

严格把控舆情方向,监测网民动态和大众感情倾向是对舆情发展控制的有力手段,也是对大数据舆论信息检测的关键所在。文章针对微博热点搜索、搜狗网页提供的微信公众号热点和百度资讯的热点新闻爬取并预测舆论倾向,利用可视化界面展示分析后的结果。首先通过爬虫获取每日热点信息,文本预处理后存入数据库;然后利用LDA主题模型提取热点事件,使用卷积神经网络分析情感倾向(正面、中性和负面);最后采用Django框架展示页面,进行相关统计研究和数据的可视化展示,通过可视化界面展示分析后得到的热点事件和舆论倾向。

Strictly controlling the direction of public opinion,monitoring the dynamics of Internet users and the emotional tendency of the public are powerful means to control the development of public opinion,and are also the key to the detection of big data public opinion information.This paper crawls and predicts the trend of public opinion for the hot search of Weibo,the hot news of WeChat public account provided by Sogou website and the hot news of Baidu News,and use the visual interface to display the analysis results.Firstly,the daily hot information is obtained by crawler,and the text is pre-processed and stored in the database.Then the LDA topic model is used to extract hot events,and the convolutional neural network is used to analyze the emotional tendency(positive,neutral and negative).Finally,the Django framework display page is used for relevant statistical research and visual display of data,and the hot events and public opinion trends obtained after analysis are displayed through the visual interface.