A Topic Detection Method Based on Microblog Weight

作者：Guo Kaijie; Shi Liang^*

来源：6th International Conference on Cyber-Enabled Distributed Computing and Knowledge Discovery (CyberC), 2014-10-10 To 2014-10-12.

DOI：10.1109/CyberC.2015.20

• topic detection
• microblog weight
• LDA

Nowadays, microblogging platforms, which have strong influence in social network, have become major platforms for Internet public opinion formation and propagation. By means of hot topic detection and topic tendency research on microblogging platforms, it will help government and companies in a timely manner to get social public opinion and popular investment spots. LDA is a commonly used method for topic detection. In order to solve the problem of describing the influence power on topic generation and propagation for different microblogs, W-LDA (Weighted LDA) model is designed to import microblog weight in LDA model. The experimental results show that W-LDA model has a better perplexity score than original model.

• 出版日期2015
• 单位厦门大学