Consensus Clustering for Simulation-Based Reputation Measurement for Online Services

User ratings for online services can be considered a categorization of the services, forming an initial clustering of them. Online service reputation measurement based on user ratings can be viewed as a further clustering of services based on the users' categorization. However, current reputation measurement methods do not consider the consistency between service reputation and users' categorization of services, leading to unreasonable outcomes. In this paper, we propose a reputation measurement method for online services based on consensus clustering. First, a user-service rating matrix is established to obtain the categorization of services by users. Next, a similarity graph is constructed using a designed similarity measurement function, which is then segmented to generate meta-clusters. Finally, the correlation between clusters and user ratings is computed to determine service reputations with maximum consistency. Simulation experiments on the MovieLens dataset are conducted to verify the effectiveness and efficiency of the proposed method.