An approximate approach for batch update of common preference group

Abstract

There are a large amount of active Web users whose behaviors reveal their preferences. The user preference model can be established via analyzing user behaviors. Mining common preference group (CPG) from the user preference model can support many applications such as potential community discovery, data sharing and recommendation. However, the user preference model is constantly evolving, which is caused by users' various behaviors. With massive users, each time the user preference model changed, mining CPG from the scratch is not an option. In this paper, we analyze how users' behaviors influence CPG. Then, an approximate approach for batch CPG update is proposed to avoid CPG re-computing from sketch when the user preference model changed. © Springer-Verlag 2013.