Context-aware ranking method for information recommendation

Abstract

Previously, we proposed two recommendation systems - context-aware information filtering (C-IF) and context-aware collaborative filtering (C-CF) - both of which are context-aware recommendation methods. We have also shown their effectiveness through of experiments using a restaurant recommendation system based on these methods. Furthermore, we need to rank the recommended contents to improve the performance of the C-IF and the C-CF. A ranking method ranks the recommended contents based on content parameters that the user regards as important. However, what parameter is important for the user is according to their contexts. For example, a user likes a reasonable restaurant when he is alone, but on the other hand, he likes an expensive and stylish restaurant when he is with his girlfriend. Therefore, it is important to consider his contexts when ranking recomended contents. In this study, we propose a context-aware ranking method. The system ranks recommended contents within users' contexts.We also evaluate our proposal method from experimental results. © Springer 2008.