Privacy Policy Fusion Based On Fuzzy Rule Reasoning

With the rapid development of the Internet era,incidents of personal privacy leaks have emerged endlessly and become more intense.Social network as a sharing platform,the resources shared on the social network may be shared by multiple sub-subjects.In order to prevent privacy leakage,each social network subject should have a corresponding privacy protection policy.Fusion of the privacy protection policies of all sharing subjects of the same content to form a unified rule expression is the key to dealing with the problem of common ownership of data.The logic rules for the privacy protection needs of different subjects are often very complex and cross each other.Once these heterogeneous rules for different privacy components are anomalous when fused,it will cause user data leakage and serious security issues.The privacy policy fusion research based on fuzzy rule reasoning proposed in this paper solves the problems of low accuracy of abstract generation rules,elimination of redundant rules and conflict rules,and the main work and research results are as follows:(1)This paper proposes an improved decision fusion algorithm based on fuzzy rules for existing fuzzy rule generation and fuzzy rule base construction methods.The algorithm has phases such as rule generation,rule evaluation,redundant rule removal,and fusion.First,a rule generation model is designed for the data source,and the effective knowledge of the data source is stored in the form of fuzzy rules.Then,considering the problems of low level of rule generation accuracy and weak convergence,the real value and coverage threshold were designed to determine the rule screening conditions and the termination conditions for building the rule base.Aiming at the problem of redundant rules in the fuzzy rule base,unnecessary rules were screened out and eliminated through the redundant rule removal method.Finally,a fuzzy rule base was established for decision fusion.The experimental results prove that the fuzzy rule decision fusion method proposed in this paper has the advantages of high accuracy and low resource consumption in most cases.(2)This paper proposes a regularized fusion method of privacy protection policies for the privacy and security issues of multi-owner data common ownership in social networks.First,according to the different meanings of the privacy data,the content of the protection object is defined.Due to the lack of a uniform description of privacy protection policies,this paper uses predicate logic formulas to abstract the natural language description of privacy protection,and further gives the logic model of privacy protection rules on this basis.Due to the different privacy protection requirements faced by multiple agents in different application scenarios,in order to ensure that these rules do not conflict and leak private data,this paper gives the definition of heterogeneous rules for privacy protection,and proposes a rule fusion algorithm based on this.Due to the different privacy protection requirements faced by multiple agents in different application scenarios,in order to ensure that these rules do not conflict and leak private data,this article gives the definition of heterogeneous rules for privacy protection,and proposes a rule fusion algorithm based on this.The experimental results confirm that the regularized fusion method of privacy protection policies proposed in this paper can fuse privacy protection policies at a higher level of rule substitution.