Research And Application Of Tolerance Granular Computing Model In Article Comparing

With the rapid development of information technology, the development of Internet specially, the source of Internet is much more plenty day by day, and then it's much more convenient to obtain information, at the same time it's much easier to copy something from others. Article Comparing System is used to compare articles automatically, and help testing staffs to search the questionable articles effectively and rapidly. Meanwhile, in the process of editing, the method of article comparing can be used to improve the editing of speed and quality. Also, it can assist editor to distinguish the problem of copyright ownership and so on. Therefore, in term of the study of article comparing have some valuable of realistic and utility.The core concept of this paper is the theory of Rough Set and Granular Computing, the method of article comparing is researched. Firstly, the basic concept of RS and GrC theory is introduced, the theory of GrC model based on tolerance relation especially. Key problems, such as the concept of GrC, the technology of chinese segmentation, text feature extraction and similarity of chinese words, are discussed. Secondly, the method of article comparing based on tolerance granular computing model is researched, and relevant arithmetic is given. Including the definition and constructing method of article tolerance granule, paragraph tolerance granule and granule space information table. Meanwhile, according to relationship of extension and complex between the tolerance granules of GrC theory, the similarity degree of article tolerance granule and paragraph tolerance granule are computed, and examples are provided. The procedure and arithmetic of article comparing based on tolerance granular computing model are elaborated on. Again, with the real data set, the arithmetic is test in different aspects, and the experimental results are analysed. In term of the experimental result, the method of article comparing based on tolerance granular computing model is effective, and satisfied with the users' requirement of finding similar articles basically. Finally, all of researches and work in the paper are summarized and prospected. The Application of Rough Set and Granular Computing in article comparing is not only extending the application of theory in the new field, but also provide a new technology for article comparing.