A topic definition model of self-media news based on Louvain algorithm
Peng Jiang, Hua Zhou, Jianhui Li · The 2nd International Conference on Computing and Data Science · 2021
In order to attract readers, most self-media writers contain keywords that reflect the theme of "net celebrity". However, the title keywords often have a certain deviation from the web page topic. In the news recommendation system, this deviation will affect the accuracy of the recommendation and thus the user experience. This paper studies the problem and proposes a self-media news topic definition model (MNLA) based on Louvain algorithm. First, determine the major topic of the news by analyzing the news headline and body tag words, then obtain the high-quality news page linked to the news, and finally determine the small topic of the web page based on the Louvain community algorithm. According to experiments, compared with other topic definition methods based on text quantity and matrix analysis, the method in this paper has certain accuracy advantages.