Research On Personalized Information Retrieval Method Based On Tags

With the development of Internet technology and the popularity of Web2.0, thesurge of network information resource indicates the coming of the era of big data. Theemergence of search engine provides a convenient for users to retrieve informationthey need in the massive network information resources to some extent. However, thetraditional search engine based on keyword matching because of the lack ofconsideration of user preferences in retrieval, it leads to that all users will get thesame results when they search the same keyword. And with the increasing amount ofinformation, the returned result set is not pithy enough, and big deviation with theneeds of users, the level of personalized service is low, and it is no longer adapt to theinformation retrieval requirements of the era of the big dada. Therefore, how toretrieve the accurate and personalized information according to the user preferencesbecame the focus of current research field of the information service.As the representative of the Web2.0technologies, tag technology has beenwidely used both at home and abroad. Aiming at the existing issues of thepersonalized information retrieval method based on tags such as semantic fuzzinessof tags, poor quality of retrieval, low efficiency of retrieval and so on, a personalizeinformation retrieval method based on tags was proposed in this paper, which solvedthe semantic fuzziness issue of tags, and improve the quality and efficiency ofretrieval, is a beneficial complementarity of the personalized information retrievalmethod, and has strong theoretical and practical significance. First proposed todescribe the network resources by using the multi-characteristic of tags, and extendedthe tags based on the extension rules and the semantic tree; Then used the improvedsimilarity algorithm and clustering algorithm to cluster user tags and selected thetopic tags, and got the user model after tags were extended; Finally, matched theresources and user model by using the multi-characteristic of tags, this paper firstmatched the their topic for preliminary screening of the resources, then matched thetags. According to the matching values, selected Top N of resources as the final retrieval results returned to the user.In this paper, first determined the background relevant theories of thepersonalized information retrieval service based on tags, which laid a foundation forthe further research work of this paper; then introduced in detail the proposeddescription method based on tags of resources and the user interests, and the matchingmethod based on tags of the resources and user model; Finally, through the contrastanalysis of experiment, proved that the proposed method had great improvement inboth quality and efficiency of retrieval. Finally, the research work of this paper issummarized, pointed out the deficiencies, and carried on the outlook for futureresearch work, proposed the further research direction.