User-sensitive Explanations under a Knowledge Pattern Lens

This paper introduces our ongoing research on a general-purpose methodology for generating explanations for occurrences of interaction patterns. Explanations are tailored around a user's profile or interaction history in interactive systems. The approach relies on the recognition of interlinks between interaction patterns and knowledge patterns, both formally modeled as networked ontologies. In addition, the method constructs its statements using the same knowledge shared by the hosting system, including distributed knowledge such as Linked Data. We plan to implement and evaluate this approach in the context of Content Management Systems and related interaction patterns such as query disambiguation, content recommendation, faceted search and browsing.