Multi Feature Extraction and Trend Prediction for Weibo Topic Dissemination Network.

In the era of big data, the extensive collection and dissemination of information has brought new security challenges, and how to ensure the security of big data under the premise of ensuring the normal operation of the topic transmission network has become an urgent problem to be solved. Online social networks have become the main channels and carriers for obtaining and disseminating information. Current social events and trending topics are transmitted in the form of topics on the microblog platform. Therefore, studying the evolution process and development trend of microblog topic communication is of great significance for public opinion monitoring, crisis prevention and control, early warning, and precision marketing. A multi-feature metric analysis method is designed for the influencing factors in the process of topic propagation, and the influencing factors are divided into two levels: content characteristics and structural characteristics. Aiming at the problem of Weibo topic trend prediction, this paper proposes a micro-blog topic trend prediction model, G-Informer, which integrates graph attention. Considering the graph structure and time evolution characteristics of the topic in the process of propagation evolution, the experimental results show that the G-Informer model in this paper has certain advantages in predicting the trend of microblog topics and has good robustness for predicting long series. [ABSTRACT FROM AUTHOR]