Research on sentiment analysis of Weibo comments based on crawlers and SVM
As an important communication platform for hot current affairs,Weibo is the focus of the attention of netizens on each article or video.Copy and paste Weibo comments after manual pulling down is daily behavior,but this operation will slow down the emotional resolution rate.For the above situations,Selenium technology is used to simulate human login and input verification codes,and import the Requests library to analyze the web source code and save Weibo reviews.Import the ChnSentiCorp emotional analysis library into the support vector machine(SVM)classification model,and after the text pre-processing of the climbing Weibo comments,the Weibo comments are classified by the trained SVM model.The classification experimental results show that the SVM classification accuracy is low.The main reason is that the emotional analysis language library is not widely available.The use of crawlers to self-built Weibo review corpus and introduce training in the classification model will make the accuracy of emotional classification higher.
Weibo commentsSelenium technologyChnSentiCorp emotional analysis librarySVMself-built Weibo review corpus
NETL
NSTL
万方数据
