Learning and exploiting context dependencies for robust recommendations
Ghim-Eng Yap · 2008
We consider the recommendation problem, where a set of available items or choices are rated and recommended to users accordingly. Over and above the ratings information used in traditional filtering algorithms, the context of the user-recommender interaction is used to improve the recommendation quality. Specifically, we study how the effective learning and exploitation of context dependencies can help to generate more personal and relevant recommendations.