Research on Sentiment Analyzer of Teaching-related Messages Based on Social Network

With the widespread use of social networking media such as Weibo, it has become a relatively popular problem in academia to analyze the sentiment tendency of these tweets through artificial intelligence. The problem of text sentiment analysis based on large-scale Weibo messages has thus become particularly important. However, there are no mature sentiment analyzers in the existing research results, and there is no research for the specific message group of Weibo (which cannot exceed 140 characters per message). This paper provides a text sentiment analysis method for Weibo messages based on social networks. A sentiment lexicon for the professional field of teaching is constructed. A large amount of Twitter message data was obtained , and these messages were classified into positive, negative and fluctuating categories after filtering irrelevant noise and a text sentiment classifier built based on Naive Bayes theory. Through testing, it is found that the method proposed in this paper can better and faster classify the emotional tendency of Weibo messages, and it is of great practical significance to use this result to guide the subsequent arrangement and decision of teaching activities, which helps to promote the construction of intelligent teaching service ecosystem.