Personalized Identity Agent for User-Centric IdM

As an emerging technology, the user-centric identity management (IdM) places a user center in which the user directly determines the offer or not when the personal information of the user is needed. But the user-centric IdM has some problems such as complicated security setting, difficulty of sharing determination of personal information, and the inefficient interaction repeat. In order to solve out these problems, we have to minimize setting procedure and present a recommendation that a user should refer to determine the sharing of personal information. It also needs to minimize user´s interaction by learning the user´s propensity. In this paper, therefore, we propose algorithms which provide a user the recommendation about the sharing of personal information in the user-centric IdM. Then we look into how much reduce the user´s interaction, and the avoidance of privacy intrusion by the changes according to the recommendation level.