Detecting user types in object ranking decisions

Abstract

With the emergence of Web 2.0 applications, where information is not only shared across the internet, but also syndicated, evaluated, selected, recombined, edited, etc., quality emergence by collaborative effort from many users becomes crucial. However, users may have low expertise, subjective views, or competitive goals. Therefore, we need to identify cooperative users with strong expertise and high objectivity. As a first step towards this aim, we propose criteria for user type classification based on prior work in psychology and derived from observations in Web 2.0. We devise a statistical model for many different user types, and detection methods for those user types. Finally, we evaluate and demonstrate both model and detection methods by means of an experimental setup. Copyright 2009 ACM.