Research And Implementation Of Public Opinion Sentiment Analysis Of University Official Accounts Based On Machine Learning

With the development of Internet technology,online life has become an important way of life for contemporary college students,and the guidance of online public opinion has also become an important part of the ideological work of colleges and universities.College students have strong ability to search and capture information,and have a high enthusiasm for participating in various topics.College students express their opinions,opinions,and emotions on school study,life,and social issues,and publish subjective texts on the online platform.The number of these subjective texts is large and growing rapidly,and it takes a lot of manpower and time to analyze these texts only manually.With the development of technologies such as natural language processing and deep learning,the use of computer technology to conduct text sentiment analysis on public opinion information in colleges and universities is of great significance.It can promote interaction and communication between schools and students,so that schools can capture students’ emotional tendencies and better understand Student thinking.In the field of text processing,deep learning can automatically learn text semantic information from a large number of samples and perform feature representation,thereby providing new ideas for sentiment analysis of public opinions in university official accounts.The main work of this paper is as follows:(1)In response to the lack of sentiment analysis data sets in the field of universities,this article refers to information such as the ranking of the activity of colleges and universities,and choose to crawl data on the two topics of school opening and postgraduate entrance examination on the Weibo official websites of Shaanxi Normal University and Huazhong University of Science and Technology.After the data is preprocessed,it meets the basic requirements for sentiment analysis of public opinion on university official accounts.(2)Aiming at the problem that the traditional neural network algorithm has weak ability to fit the global information of the text and the local key features cannot be better extracted,the BiLSTM neural network is used,and the self-attention mechanism is introduced to increase the attention to the partial key information of the text.The BiLSTM sentiment analysis algorithm based on the self-attention mechanism has been compared with SVM,CNN,BiLSTM and other algorithms to obtain better experimental results.(3)The sentiment analysis algorithm of BiLSTM based on the self-attention mechanism designs and constructs a sentiment analysis system for public opinion in colleges and universities,judges the sentiment tendency of college text data,and realize the functions of viewing university public opinion information of different schools and different topic categories.