Automating the Discovery of Recommendation Knowledge

In case-based reasoning (CBR) systems for product recommendation, the retrieval of acceptable products based on limited information is an important and challenging problem. As we show in this paper, basic retrieval strategies such as nearest neighbor are potentially unreliable when applied to incomplete queries. To address this issue, we present techniques for automating the discovery of recommendation rules that are provably reliable and non-conflicting while requiring minimal information for their application in a rule-based approach to the retrieval of recommended cases.