Personalized Search System Of Internal Research And Realization

Influenced by the upgrading of information systems and the booming of web 2.0 applications,the information inside a modern enterprise is growing explosively,and lowers the efficiency of employees and capability of getting useful information.Therefore,enterprise search is a hotspot in the information retrieval area.In enterprise search,there are two key issues:what to search and how to search.As for the content to be searched,there are related researches and practices to help abstract and format the huge data which are unstructured and dispersive.This thesis focuses on how to search,which means how to help enterprise user find the most needed information within the huge data.Personalized search technology is born for this problem.Nowadays personalized search technology is mostly applied in the Internet.However, because of the security and privacy problems in the Internet,the steps of developing this technology are slowed down.In contrast,the intranet in the enterprise is safe and reliable.Furthermore,the enterprise is always encouraging knowledge sharing.Under these circumstances,employees will open and share their interests and knowledge with pleasure,which is favorable for the application of personalized search technology.This thesis is to describe many models involved in the personalized search,including user interest model,classic filtering model and search model.After analyzing the short points in the classic filtering model,this thesis creatively puts forward the role-based collaborative model and design an enterprise personalized search system based on the new model.The prototype system is integrated with a web 2.0 knowledge sharing platform,and gets a good result.