Discovering Topical Interactions in Text-based Cascades using Hidden Markov Hawkes Processes
This addresses the gap in modeling topic interactions in social media conversations for researchers and practitioners in network analysis and information diffusion.
The paper tackled the problem of modeling topical interactions in social media cascades, proposing the Hidden Markov Hawkes Process (HMHP) to jointly capture topic-topic, user-user, and user-topic patterns, and showed that it generalizes better and recovers network strengths, topics, and diffusion paths more accurately than state-of-the-art baselines in experiments on real and semi-synthetic data.
Social media conversations unfold based on complex interactions between users, topics and time. While recent models have been proposed to capture network strengths between users, users' topical preferences and temporal patterns between posting and response times, interaction patterns between topics has not been studied. We propose the Hidden Markov Hawkes Process (HMHP) that incorporates topical Markov Chains within Hawkes processes to jointly model topical interactions along with user-user and user-topic patterns. We propose a Gibbs sampling algorithm for HMHP that jointly infers the network strengths, diffusion paths, the topics of the posts as well as the topic-topic interactions. We show using experiments on real and semi-synthetic data that HMHP is able to generalize better and recover the network strengths, topics and diffusion paths more accurately than state-of-the-art baselines. More interestingly, HMHP finds insightful interactions between topics in real tweets which no existing model is able to do.