A Taxonomy for Generating Explanations in Recommender Systems

This article proposes a taxonomy to categorize and review the research in the area of explanations. It provides a unified view on the different recommendation paradigms, allowing similarities and differences to be clearly identified. Such information is commonly exchanged between a sales assistant and a customer during in-store recommendation processes and is usually termed an explanation (Brewer, Chinn, and Samarapungavan 1998). We define explanations in recommender systems by two properties. First, they are information about recommendations, where a recommendation is typically a ranked list of items.