Belief revision for adaptive recommender agents in E-commerce

Abstract

Recommender systems are a well-known technology for E-commerce. This paper illustrates a novel recommender agent model which combines the inductive power brought by text mining methods with the non-monotonic reasoning capability offered by a belief revision system to improve the agents' prediction effectiveness, learning autonomy, and explanatory capabilities. © Springer-Verlag 2003.