Research On Privacy Policy Management And Enforcement In Social Network

Social Network Services are more and more popular and a large number of users register on these websites,where they share personal information and communicate with others.They are convenient to our lives, but they may be cause personal privacy disclosure. Therefore, how to protect the privacy of social network has become a hotissue.Policy-based access control is one of the methods to protect privacy in Social Network.Policy-basedprivacy management mainly includes policy definition, validation and enforcement. The definition is formal specificationaccording to users’ privacy preferences, which is the basisof enforcement. The verification is used todetect policy conflicts to ensure the policy’s consistency. The enforcement is the fine-gained access control, which is the concretemeans of privacy protection. This paper researches on privacy policy management forSocial Network. It main contributions include:Ⅰ.We propose the personalized privacy policy definition and implementation model. It integrates policy definition, validation, recommendation and enforcement.The privacy policy definition is used todescribe users’ privacy preferences in the first-order logic language. It realizes role authorization to unknown users, as well as permission assignments to dynamic resources through rule inferences.Ⅱ.We put forward logic consistency validation to analyze privacy policy conflicts.Thefirst we translate the privacy policy information into Prolog facts, and then design the policy consistencyinference rules, later query Prolog target based on the facts and inference rules to find instances of the policy conflicts.Ⅲ.We propose personalized privacy policy recommendation model. It analyzes the relationships between groupsand attributes as well asbetween permissionsand data labels. Weuse information entropymethods and association rulesto mine users’privacy preferences from users’ history. Because the data label is a natural language with ambiguity and uncertainty, we process data labels’ semantic analysis to enhance the accuracy of privacy policy recommendations.Ⅳ.We design and develop the personalized privacy middleware system for Social Network. It achieves privacy policy setting, policy conflict analysis, and fine-grained access control. It verifies the feasibility of the personalized privacy policy definition and implementation model. Also we analyze the efficiency of policy consistency verification by experiments.