An Analysis of Emotion and User Behavior for Context-aware Recommendation Systems using Pre-filtering and Tensor Factorization Techniques
The emotion-based context-aware recommendation systems have been widely adapted by a wide variety of recommendation domains despite the fact only a few studies have been analyzed in tourist destination recommendations. To utilize the concept of user emotion and incorporate it into the recommendation process along with user behavior, we proposed a travel destination recommendation system. Also, we compare and clarify the effectiveness of using emotion and user behavior in the recommendation process by suggesting a framework based on two techniques: filtering and contextual modeling. For the filtering based approach, we used Prefiltering, and for the contextual modeling, we employed Tensor Factorization. Both these approaches performed excellently with the selected contexts in the proposed framework, and the results of the Tensor Factorization approach proved to be highly effective in tourist destination recommendation compared to other Pre-filtering.
