Multi-Aspect Collaborative Filtering based on Linked Data for Personalized Recommendation
Han-Gyu Ko, Joo-Sik Son, In‐Young Ko · 2015
Since users often consider more than one aspect when they choose an item, relevant researches introduced multi-criteria recommender systems and showed that multi-criteria ratings add values to the existing CF-based recommender systems to provide more accurate recommendation results to users. However, all the previous works require multi-criteria ratings given by users explicitly while most of the existing datasets such as Netflix and MovieLens are a single criterion. Therefore, to take advantage of multi-criteria recommendation, there must be a way to extract necessary aspects and analyze users' preferences on those aspects from the given single-criterion type of dataset. In this paper, we propose an approach of utilizing semantic information of items to extract essential aspects to perform multi-aspect collaborative filtering to recommend users with items in a personalized manner.