LAUSR

Content-Based recommender systems (CB) filter relevant items to users in overloaded search spaces using information about their preferences. However, classical CB scheme is mainly based on matching between items descriptions and user profile, without considering that context may influence user preferences. Therefore, it cannot achieve high accuracy on user preference prediction. This paper aims to handle context-awareness (CA) to improve quality of recommendation taking contextual information as the trend in current trend interest, in which a stream of status updates can be analyzed to model the context. It proposes a novel CA-CB approach that recommends question/answer items by considering context awareness based on topic detection within current trend interest. A case study and related experiments were developed in the big data framework Spark to show that the context integration benefits recommendation performance.