Research And Realization Of Privacy Policy Generation System Based On CBR And Ontology

In recent years, with the great popularization of wireless networks, mobile devices and localization equipments, pervasive computing has rapidly become a usable entity and a part of daily lives. Lots of advanced services have merged that are developed to meet this trend. On the one hand the users have enjoyed a wide range of services, but on the other hand they are bothered by the inevitable user privacy pitfalls.Managing security and privacy policies are known to be a difficult problem. Lay users often do not understand their own policies. In mobile and pervasive computing settings; this situation is often exacerbated by its characteristic of distribution and heterogeneity. To make matters worse, the users have to engage in numerous tasks concurrently. So the privacy policy management in pervasive computing is more complex than in traditional computing environments. It places a larger burden on users and .lower users' satisfaction with the services then hinders the further promotion of Pervasive Computing.CBR (Case-based reasoning) is a new artificial intelligence method, it solves new problems by retrieving records of similar past problem solving episodes and adaptation the prior solutions to fit the current situation.This paper first makes a study of the principle, basic process and the core technology of case-based reasoning. Then it introduces the definition, classification and description language of ontology, as well as the criteria, methods and tools for creating ontology and so on. This paper also introduces the design objectives and overall structure of this privacy policy generation system.This paper builds up the Privacy Policy generation system based on J2EE framework. This system can retrieve the default privacy policy for user based on the basic user information and can greatly improve the efficiency that users articulate the policy. This article introduces the ontology concept into the CBR case retrieval process. The case-specific knowledge and corresponding domain knowledge are combined to enrich the knowledge of system. The use of ontology makes the original isolated discrete values become the instances with semantic relations. Then this paper forwards the concept of ontology path and ontology similarity to improve the accuracy of the retrieval. On the premise of guaranteeing similarity, in order to provide more choice to users, this paper has proposed the improved KNN algorithm based on diversity theory that can better satisfy users. Based on the analysis of user's basic information, this paper extracts user's effective information to make up case's description field and establishes the expression form of the case. In regard to the construction of the case base, the paper has achieved the double layers storage structure through using Hibernate. This structure can deal with heterogeneous databases and offers the system great flexibility and portability. Finally, this paper uses version control system to manage privacy policy data and achieves the version storage of policy that can provide users some convenient and helpful functions such as showing historical information; this has provided a new way of thinking for the data storage of information systems.Finally, the paper introduces the system interface design and summarizes the work that shall be done in future.