Tracking Aspects in News Documents

摘要

In this investigation, we discuss aspect tracking, i.e., how to identify tracking story-lines of document topics. Because a huge amount of fragment information occurs, it is difficult to see what they mean and how they go within topics by hand. Here we are addressing this type of problem using stochastic models. Our basic idea is that we consider state transitions as internal structure of stories based on HMM and we extract some story-lines as aspects of topics by probabilistic likelihood. We will use KL divergence to screen topics. Also, we discuss topic detection, i.e., how to construct clusters of events to discuss the same topic.