Handling Topic Drift for Topic Tracking in Microblogs

摘要: Microblogs such as Twitter have become an increasingly popular source of real-time information, where users may demand tracking the development topics they are interested in. We approach problem by adapting effective classifier based on Binomial Logistic Regression, which has shown to be state-of-art in traditional news filtering. In our adaptation, we utilize link information enrich tweets’ content and social symbols help estimate quality. Moreover, find that very likely drift microblogs a result redundancy topic divergence tweets. To handle over time, adopt cluster-based subtopic detection algorithm identify whether occurs detected is regarded current focus general adjust drift. Experimental results corpus TREC2012 Microblog Track show achieves remarkable performance both T11SU F-0.5 metrics.