Network Public Opinion Monitoring Method Based on Machine Learning

To ensure the public safety of social networks, this paper designs and implements a public opinion monitoring system based on Sina Weibo. A design scheme based on machine learning and word segmentation technology is proposed. Jieba Chinese word segmentation tool is used to realize text preprocessing, text classification and tendency analysis. When classifying according to the emotional weight of emotional words in the text, LSTM model is used to improve the analysis efficiency and accuracy of machine learning. At the same time, the web page library update strategy for the public opinion analysis system is formulated to ensure the freshness of the local web page library. The experimental test analysis shows that such scheme can classify the emotional tendency of microblog content, and find out the news with high attention and strong emotion of netizens, and make the corresponding trend prediction, which achieves the purpose of public opinion monitoring.