Large Group Failure Mode and Effects Analysis Method Considering Individual Concerns and Integrated Consensus Mechanism in Social Network Context: An Application to Mobile Medical Privacy Risk Evaluation

With the popularization and application of mobile medical applications (apps), the privacy and security of personal medical data has received widespread concerns. Given that privacy leaks can cause serious privacy violations and personal safety issues, it is particularly important to adopt failure mode and effects analysis (FMEA) in privacy risk assessment to ensure reliability and security. In particular, FMEA risk assessment requires the participation of numerous experts from different fields. Experts often express opinions with a level of individual concerns and confidence levels that needs to be considered in the risk assessment process. In this paper, a large group FMEA considering individual concerns and integrated consensus mechanism in the social network environment is developed for privacy risk assessment of mobile medical apps. Firstly, this paper constructs a complete social trust network (STN) based on the trust propagation operator proposed by the trust propagation path, which is conducive to comprehensively analyzing the social relationship among experts. Secondly, a hierarchical clustering method is proposed that considers both opinion similarity and concerns similarity between experts. This clustering method processes information from both subjective and objective perspectives to ensure a more reasonable cluster division. Next, considering that the level of consensus within a cluster affects the efficiency of consensus reaching, an integrated consensus mechanism that considers both intra-group consensus and inter-group consensus is developed to promote the unification of opinions. Additionally, this paper utilizes social trust relationships to guide the adjustment of the consensus level to reduce information loss on the premise of satisfying the expert's confidence level. Finally, a case study on assessing privacy risks of mobile medical apps illustrates the effectiveness of this approach, and the superiority and rationality of the method are confirmed by multi-angle comparative analysis, simulation analysis and parameter analysis.