SigmoRep: A robust reputation model for open collaborative environments

Abstract

In recent years, numerous collaborative web applications have been successfully used to create encyclopedias, learning-objects, maps and even artworks. Trust and reputation systems are two important components of all collaborative information repositories such as Wikipedia. The open nature of such repositories encourages users to contribute to the application. However, this also makes vandalism and low quality contribution possible. Reputation systems try to provide some mechanisms to evaluate the quality of contributions and prevent malicious and untrusted users from harming the resources and encourage constructive behaviors. In this paper, we present an innovative reputation system named SigmoRep, which uses article evaluation patterns to calculate the author's reputation values. It uses a function called controlled influence function to prevent oligarchy. In this paper, we also investigate the evaluation of Wikipedia's articles at the fragment level. Our reputation system is resistant to a wide range of related attacks and is more human-understandable than other similar systems.