Towards Time-Dependant Recommendation based on Implicit Feedback

This work introduces a new context-aware recommendation approach called user micro-proling, which split each single user prole into several possibly overlapping sub-proles, each representing users in particular contexts.

Abstract

Context-aware recommender systems (CARS) aim at im-proving users ’ satisfaction by tailoring recommendations to each particular context. In this work we propose a con-textual pre-filtering technique based on implicit user feed-back. We introduce a new context-aware recommendation approach called user micro-profiling. We split each single user profile into several possibly overlapping sub-profiles, each representing users in particular contexts. The predic-tions are done using these micro-profiles instead of a single user model. The users ’ taste can depend on the exact partition of the contextual variable. The identification of a meaningful par-tition of the users ’ profile and its evaluation is a non-trivial task, especially when using implicit feedback and a contin-uous contextual domain. We propose an off-line evaluation procedure for CARS in these conditions and evaluate our approach on a time-aware music recommendation sytem. 1.