Research And System Implement Of Multi-agent Personality Information Service System Based On Web Mining

The internet is an open, dynamic, isomeric network, in which resource distribution is dispersed, information quantity is very large, and lacking unified management. Users have to spend a lot of time to finding information. Search Engine and variety professed information retrieval systems lighten the burden on the people. But nowadays neither these search engines nor the professional retrieval system can not satisfied the user's various desire, and we must spend so much time in filtering the information we get, so the conception of personality information service is be mentioned.We have researched the knowledge of personality information service, and then proposed and implement multi-agent personality information system, which is based on web mining, for Jiangsu Engineering Technology Literature Information Center information service system. The system is composed of three Agents: user agent, information filtering agent, information recommend agent. The Three Agents work dependently and have contact with each other. User Agent is in charge of user model building and the interaction of user-machine interface. User model building includes user information acquisition, the representation of user model and the update of user model. Information filtering agent is in charge of information filtering, using content filtering and collaborative filtering. Information recommending agent will recommend information to users, and our system supply web recommending and email recommending. The Three agents is group work, and complete the personality information service together. The Three adopt coordinate structure and work flexible.This System is the first personality information system using TRIP system's query log for web mining at home. User Agent uses TRIP system's query log to extract users' interest, and build users' model. And then we bring in date parameter based on TF-IDF when we compute users' interest index.At last The system is tested based on TRIP system's query log. And the test includes performance of process and the performance of recommending.The first and second chapters are overview about personality information service and the technology related. The Chapter 3 described the theory model of multi-agent personality information service system based on web mining, including their relevant knowledge and architecture. The Chapter 4 introduced the implement of multi-agent personality system based on web mining, including background introduce, implement process and system...