DOI:
Abstract:
We assume that summaries vary according to the users and factors such as their information needs, situations and purposes in seeking information, so that different summaries can be produced according to the author’s point-of-view. In our summarization system, a list of topics discussed in the given document set are presented, so that the user can select a topic of interest from the list as well as the summary type, such as opinion-oriented, fact-reporting or data-centered. The purpose of this study is to build a multi-document summarizer to produce summaries according to viewpoints. We assume a relationship between summary type and genre because both features are orthogonal to topics. As an exploratory stage of investigation, we examined the effectiveness of genre for source documents to produce different types of summaries. We also analyzed the results of the multi-document summarization using automatic genre classification to reveal the association between genre dimensions and the summary types. Our experiments showed that the "argumentative" dimension of the genre of the source document was an effective feature to discriminate multiple document summaries and had effects on "opinion-oriented" summary types.