A Study On User Ontology Profile Based On Formal Concept Analysis

Personalized information service is based on the theory of knowledge management and has been the main method to solve problems such as 'Information-mazing' and 'Information Overloading'. Personalized information service is based on good learning and accurate description of the user. As a general description of the user, user profile is key to the personalized information service and decisive to its efficiency and quality. User ontology profile is a new form of user profile. Compared with traditional user profiles, it gives a general and accurate description of the user and has more abundant semantic intension. Currently most researches on user ontology profile pay more attention to the description of the user's characteristics. And there is a lack of studies on the automatic building of user ontology profiles. So it is necessary to propose a user ontology profile suit to be built automatically and to study the corresponding automatic building method.For this, after studying personalization system, the theory of Formal Concept Analysis (FCA) and the characteristics of several user profile and corresponding building methods, a general architecture of user personalization is proposed, a user ontology profile in the form of weighted concept lattice is automatically built, a user personalized prototype system is designed, and then the user ontology profile and its building method is analyzed and evaluated with an arithmetic case study.The main work of this dissertation is as follows. At first, the general architecture of user personalized system is proposed according to the system analysis. And then based on Formal Concept Analysis, the website domain ontology and user ontology profile are defined and the corresponding arithmetic is designed. What's more, a prototype of user personalized system is designed and the user ontology profile and its building method is analyzed and evaluated with an arithmetic case study. In the case study, users of an experiment website are studied and a manually built user ontology profile, a user ontology profile in the form of concept hierarchy and our user ontology profile are given to each use, then compared and analyzed.As far as theory is concerned, this research is a good supplement to the researches on the form and automatic building of user ontology profile. In practice this study can be used in personalized information service to improve its efficiency and quality, and then contribute to its development.