An Improved Latent Dirichlet Allocation Model for Hot Topic Extraction

摘要: Micro blogging is fast becoming a dominant medium in social media and its impact evident our daily lives. A massive amount of information produced on basis. It observed that detecting hot topics can be very helpful for people to get essential quickly. But due short sparse features, high flood meaningless tweets other characteristics micro blogs, traditional topic detection methods are unable achieve desirable level performance. In this paper, we propose multi-attribute latent dirichlet allocation (MA-LDA) model, analysis model which the time tag attributes blogs incorporated into LDA model. By introducing variable about attribute, MA-LDA decide whether word should appear or not. Applying attribute allows rank core words results so expressiveness outcomes improved over Empirical evaluation real data sets demonstrate method able detect accurately efficiently with more terms associated each found. Our study provides strong evidence importance temporal factor extraction.