TextRank enhanced Topic Model for Query focussed Text Summarization

Topic model based query focused text summarization methods can meet the user's needs as they are able to find subtopics and their correlations. Topic model based summarization method does not score the sentences directly as they generates topics distributions, which are used to score the sentences. Graph based methods rank sentences using syntactic and semantic information. In this work, a topic model based summarization method namely two-tiered topic model is combined with graph based TextRank method. The combined method, called TextRank enhanced Two-Tiered topic model, uses the important sentences obtained from TextRank in the generative process of two-tiered model to extract better summary sentences. The proposed method's summary results outperform other topic model based summary results using ROUGE metrics evaluated on DUC 2005 dataset.