Enhancing Context-Aware Recommendation via a Unified Graph Model
Hao Wu, Xiaoxin Liu, Yijian Pei, Bo Li · 2014
In order to assist the development and use of context-aware recommendation capabilities, we propose a unified graph model which incorporates contextual information in an advantageous way. Basic RWR and Context RWR are specifically designed to estimate the relevance between the target user and the items against the contextual graph. Also, we propose two post-processing strategies to filter recommendation results given contextual conditions. Being dependent on experimental results on two datasets, all proposed methods are effective to enhance the quality of context-aware recommendations.