| In the fast developing age of Information Technology, People whose life is fast-paced can obtain precise useful information with very simple methods in the internet. In recent years, search engines are developing rapidly. The desktop search engines have also been developed rapidly but personalized services have not satisfied users'requirements. They are simply locating documents. Analyses of knowledge contained in users'documents, mining of potential knowledge and search results exhibition with humanization are their defects. These points are the focuses of this thesis.The thesis stems from one of the Sichuan Provincial Youth Fund for software innovation projects—U-CLASS. The platform constructed in this thesis intends to provide users with knowledge mining and management, personalized search, users'interests learning and personalized service and etc. The work of this thesis is described as the following:(1) From perspective of basic concepts and history of search engine, the development and advantages of desktop (document) search engine are analyzed. The research direction is proposed with U-CLASS;(2) Clustering based on parallel fuzzy ants and some other algorithms are proposed; A large amount of data and some testing tools are used to assure accuracy and high efficiency of models and algorithms;(3) More detailed work of design and implementation is done; top-down approach is mainly used; from the perspective of system structure of each module and package design, the thesis has a clear structure. The main module or algorithms are as the following: the clustering algorithm design and implementation and etc.;(4) A number of tools related to software testing are used to do functional testing, pressure testing and etc.; JProfiler is used for performance analysis and optimization of the whole system;(5) The follow-up work to be done is concisely analyzed and prospects are presented while the work accomplished in this thesis is summarized. |
PDF Full Text Request