Research On Topic Hotness Trends Prediction Of College Network Public Sentiment

With the development of computer and network technology and the popularity of mobile terminals, the Internet platform has become the fourth media after newspaper, radio, television and other traditional media. On the Internet platform, users can publish their views and opinions, fully expressing their wishes. However, it's not all that unusual that one or several individuals' expression of opinion will spark a huge wave of public sentiment, which in many cases could point to or even cause certain social problems.The thirty-ninth China Internet Development Statistics Report published in January 2017 shows that college students have become a large and influential proportion of Internet users because of their great number and active thinking. College students are willing to put their own emotions, attitudes, opinions etc. on the Internet platform. The rapid spreading of the students' personal opinions on the Internet, once having aroused widespread concern, could evolve into group events. This has certain influence on college campus and college students' thought. On the National Conference on Ideological and Political Work in Colleges and Universities in December 2016, General secretary Xi Jinping emphasized the importance of making use of new media as well as new technologies in the ideological and political work in colleges and universities, and further exploiting the positive effects of Internet technologies on serving and ideologically guiding students in colleges and universities. Therefore, research on topic hotness trends prediction of college network public sentiment, as a scientific way of providing guidance to the educational, administrative and serving work of colleges and universities, holds great theoretical and practical significance to colleges and universities in keeping track of the thoughts of their students, providing students with ideological and political guidance and improving the timeliness and pertinence of their administrative work.This paper analyzes the connotation, structure and characteristics of college network public sentiment, as well as its positive and negative influences, and divides the college network public sentiment media into controllable media and uncontrollable media. The paper then describes the topic definition, forms of expression, life cycle, communication classification and hotness of college network public sentiment, puts forward the topic hotness trends prediction framework of college network public sentiment and puts forward topic hotness trends prediction model based on LSTM. Finally, the paper takes the "sea and sky" forum as an example, finds topics of college students' attention on forum through K-Means clustering, calculates topic hotness and predicts topic hotness trends in order to verify the rationality and superiority of the model.