Research On Query-directed Multi-mode Automatic Summarization

Along with the rapid development of Intemet,information acquisition and utilization are convenient.However,it also brings the problem of information overload.How to extract information quickly and accurately from massive information in Internet becomes more and more difficult.The goal of Query-directed Multi- document Summarization is to epurate content from the returned document set and to provide a brief description to users.The description should be relative to the given query and can satisfy the individual requirements.It can significantly improve the efficiency of information acquisition and utilization.In this paper,we design a variety of summary mode for Query-directed Multidocument Summarization to satisfy the individual requirements.The research work and innovation are as follows.Firstly,the theory and some methods of complex network are successfully. introduced to the field of automatic summarization.The documents are represented as two-layer complex networks,whose nodes describe text and paragraph respectively. The incremental modular degree matrix is endowed with.weight,and the thought of CNM algorithm is used to cluster text and paragraph.Secondly,the strategy of summary element extracting is presented.A sentence scoring method based on,keywords extracted is used to extract sentences.Then,the information density of every sub-topic is measured,and the results are used to decide how many sentences will be extracted from the corresponding sub-topic.Besides,the strategies of node importance of complex networks are taken to mine the core paragraphs as another summary element.Thirdly,on the basic of network structure of documents,we design four summary modes besides the traditional summary mode.They are document summary,general summary,partial summary,global summary and detailed summary.With the clues of sub topic,Users can browse information in certain logical sequence to their own.