Policy Generation for Privacy Protection Based on Granular Computing

Privacy is an increasingly cited concept in the large-scale deployment of web services in pervasive computing. The necessity of putting data-owners' personal information online and unwanted intentions to access data-owners' information augment the risk of privacy disclosure. To address this problem, we propose a policy-based privacy protection mechanism that protects personal information by policy from privacy violations. The policy is automatically generated based on granular computing, which attains policies about information disclosure degree from the historical access records, and data-owner's feedback improves the accuracy of our algorithm. Due to the obvious advantages to describing problem spaces at different granularities and hierarchies in granular computing, we present Infospace, a framework for storage and description of personal information in different predefined hierarchies. Besides, through case study we show the effectiveness of our new mechanism.