A model of a virtual community with a decentralized reputation-based peer evaluation

Abstract

This study was motivated by the problem of identifying fake documents on the Internet. To explore possible solutions to this problem we introduce a model of a network community in which members submit documents with verifiable content. Documents are evaluated by three individuals, randomly selected from all the members. The members whose documents are evaluated as authentic gain reputation. Otherwise, they lose it. Members with higher reputation have a greater chance to be selected to serve as evaluators. Analytical and computational results suggest this evaluation mechanism is effective in a wide range of parameters including the presence of members who consistently submit fake content and do not evaluate documents of others objectively. The proposed mechanism can be used in a variety of situations including evaluating photo and video files with geolocation and time data and creating collections of digital evidence of events clear from fake materials.