Movatterモバイル変換

Agent privacy in social networking systems is directly affected by many factors, mainly agent interaction and social activities such as creating relationship links and content sharing between agents. Based on these activities, agent privacy in social networking systems can be protected at three access control levels that are at individual, social and content levels. However, an agent may unintentionally violate its privacy during runtime by overexposing or oversharing its personal, relationship or content information due to the dynamic of the agent privacy preferences and social behavior. Hence, a solution to monitor the level of agent privacy during runtime is needed. In this research, a solution called Runtime Verification and Quality Assessment (RVQA) is proposed to improve agent privacy violations detection during the execution of agents social activities by combining the verification and assessment process from individual, social and content levels. New agent privacy requirements, parameters, checking rules and algorithms to detect agent privacy violations are presented. The effectiveness of the proposed solution is evaluated by implementing RVQA within agent-based social networking system model.