Public Opinion Analysis of Weibo Comments Based on Crawler and SVM

Abstract

In order to predict the trend of public opinion in Weibo news, this paper proposes a public opinion analysis method based on the combination of crawler and SVM. Firstly, word2vec model is used to train the sample, and the results are used to train SVM. Then, according to the popular news comments on Weibo, crawler is used to get the news, Jieba is used to segment the words into the model to judge, and hierarchical vector machine is used to judge the emotion. Finally, based on the statistical data to judge the trend of public opinion. The experimental results show that the test result is relatively more accurate and effective for public opinion analysis.