A Joint Model of Conversational Discourse and Latent Topics on Microblogs Jing Li∗ Tencent AI Lab Shenzhen, China
[email protected] Yan Song Tencent AI Lab Shenzhen, China
[email protected] Zhongyu Wei Fudan University School of Data Science, Shanghai, China
[email protected] Kam-Fai Wong The Chinese University of Hong Kong Department of Systems Engineering and Engineering Management, HKSAR, China
[email protected] Conventional topic models are ineffective for topic extraction from microblog messages, because the data sparseness exhibited in short messages lacking structure and contexts results in poor message-level word co-occurrence patterns. To address this issue, we organize microblog mes- sages as conversation trees based on their reposting and replying relations, and propose an unsupervised model that jointly learns word distributions to represent: (1) different roles of conversational discourse, and (2) various latent topics in reflecting content information. By explicitly distinguishing the probabilities of messages with varying discourse roles in containing topical words, our model is able to discover clusters of discourse words that are indicative of topical content. In an automatic evaluation on large-scale microblog corpora, our joint model yields topics with better coherence scores than competitive topic models from previous studies. ∗ Jing Li is the corresponding author. This work was partially conducted when Jing Li was at Department of Systems Engineering and Engineering Management, The Chinese University of