Research On A Model Of User Profile Based On Concept And Semantic

It is well known, the rapid popularization of Internet, it can promote the global informationization construction, at the same time, are also taking place of profound changes in lives and thinking mode of people. Internet as an indispensable source of information, because it lacks a unified organization and management, the internet resources is unorderly, web site quantity and content changes in the moment, it has caused the contradictions between excessive information and the limited information acquisition ability of people, the question of "the information to overload" and "the information confused approaches" was appeared.So People proposed personalized information services, personalized information services took one recent information service way can improve resource utilization and access efficiency, meet the human's personalization needs. Along with personalized service development and research thorough, people are becoming aware that the quality of personalized service does not only depend on recommendation and retrieval technology, but also depend on the computability description of user interest and preferences, the latter is more important. User interest model as the basis and core of personalized service has become a hot research topic, but it has not been a clear standard. This paper do an in-depth study in various modeling method involve in the source of information to obtain, description, interested size, updating mechanism, sharing and reuse, these has advantages and disadvantages.In this paper, on the basis of in-depth understand the user interest model involving in artificial intelligence, machine learning, neural networks, rough set theory, data mining, and many other aspects and fields, analysising the current model of user interest in the form of access, retrieval, description, update, evaluation to information etc., making the use of concept semantic space and ontology build model based on the concept semantic and modeling method. Model uses a combination form of concept level tree and concept set describe user interest model. In the process of model building, first of all, according ontology and china library taxonomy to establish the concept level tree, then use the demonstration to study to generate standardized expression of branch of the concept level tree, this kind of standardized expression is indicates by the concept set. Through excavating concept and the concept semantic relations in the user feedback documents, catches and expresses the user the interest. The experiment indicating: comparing with the traditional vector space model based on keywords, model based on semantic concept can positioning to users interest information more accurately, having better sharing and reuse,can enhance the the performance of user interest model.