Research On Individuation Information Search Service In Client-side

With the explosive growth of Web information,how to quickly and accurately find the necessary information from the vast information resources has become a major challenge. Though the traditional search engine technology meets the people's needs to a certain degree, however, because of its universal nature, it still can not meet the users'individuation needs of the different backgrounds, different purposes and different interests. The individuation information search service aiming at this problem has been proposed.Individuation information search service for different users refers to the different characteristics of services provides different service strategies and contents. Individuation information service includes the client-side form and the server-end form, this paper mainly studies the individuation information service of client-side form.This paper shows the common search engine system's structure and the working process, elaborates key technologies of realizing a search engine, proposes the developing process of realizing the individuation information service,discusses the definition of individuation information search service, its classification and characteristics. The construction of client-side individuation search engine and individuation searching algorithm are presented.Generally speaking, there are two ways to gather user's interests: passive and active. The individuation search based on template proposed by this paper unified these two ways. And the realizing of establishing the initial user's interests description belongs to the passive way, which mainly gathers user's interests information by template's information inputs, and the original interests model is obtained. The realizing of collecting users'feedback information belongs to the active way, which does not need the user to input the interested information personally, but actively discover the users'interests from the users'usual network browsing customs, and then the users'interests model can be further optimized. The personalized search based on the users'implicit expression information also belongs to the active way.According to current individuation information service classification, this paper proposes a three-level users'interests'structural model aimed to LAN. The model means the individuation process simultaneously processed both on the client-side and the sever-end, this may let the user experience the individuation information service more perfect.This paper proposes new methods of describing the users'interests to improve the tuple's vector description of user's interests. The first kind describes the users'interests with the forest structure, the various aspects of users'interests will get a more reasonable description. Another kind is the multistage structure which expands the three-level interests'structural model discussed above.Finally, this paper has developed a full text search engine experiment prototype system based on Lucene by using the Java language on the Windows platform. Based on this platform we has realized several kinds of individuation information search service proposed in this paper. A news search system sorted according to the date implemented by this paper has realistic significance, for no such sorting rule adopted by any news search system at present. In the ending part of this paper we summarized the whole text and made further research work's tentative plan.