摘要
新闻是网络舆情的重中之重,网络上的新闻种类繁多。当出现重大负面新闻时,若不能及时了解和引导,很容易形成舆论危机,严重时甚至影响公共安全。基于微服务架构,使用爬虫搜集新闻;利用TF-IDF提取关键词,采用Simhash算法去重;建立深度神经网络模型对新闻的类型与情感倾向进行分析。最后将这些模块连接成系统,对新闻内容进行分析预测并在前端展示,以期能及时监测到负面新闻,便于迅速采取必要措施。测试结果表明,系统是稳定可行的。
News is the top priority of public opinion,for there are many kinds of news on the Internet.When there are major negative news,if we can’t understand and guide them in time,it is easy to form a public opinion crisis,and even affect public security in serious cases.This paper used microservice architecture,crawler to collect news,TF-IDF to extract keywords,and simhash algorithm to de-duplicate news,established deep neural network model to analyze news types and emotional tendencies,and finally connected these modules into a system to analyze and predict news content and displayed it in the front end,so as to monitor negative news in time and take necessary measures.The tests show that the system has completed the expected function,and as the core part,it is applied in a securities company project to monitor the daily dynamics of listed companies.
作者
周雁
王庆娟
庞桐
Zhou Yan;Wang Qingjuan;Pang Tong(School of Computer Science and Technology,Zhuhai College,Beijing Institute of Technology,Zhuhai 519000,China)
出处
《台州学院学报》
2021年第3期13-19,共7页
Journal of Taizhou University
