The power of suggestion

Abstract

User feedback is vital in many recommender systems to help guide the search for good recommendations. Preference-based feedback (e.g. "Show me more like item A ") is an inherently ambiguous form of feedback with a limited ability to guide the recommendation process, and for this reason it is usually avoided. Nevertheless we believe that certain domains demand the use of preference-based feedback. As such, we describe and evaluate a flexible recommendation strategy that has the potential to improve the performance of case-based recommenders that rely on preference-based feedback. 1