Analysis of Information Dissemination Orientation of New Media Police Microblog Platform

This article aims to analyze the microblog data released by a new social media platform for police in a certain province in China, and find out the characteristics of those microblogs which are more likely to be reposted from the perspective of modern police affairs. A new topic-based model is proposed in this article. In the first step, the LDA topic clustering algorithm is adopted to extract the topic categories with forwarding heat from the most-reposted microblogs, with a Naive Bayesian algorithm as followed to predict the topic categories. The pre-processed sample data is applied on the proposed method to predict the type of microblog reposting. The experimental results utilizing a large number of online microblog data demonstrate that the proposed method can accurately predict the reposting of microblogs with high precision and recall rate.