Abstract
Recommender systems (RS) are the most widely used technique to suggest the users their likings based on the ratings and ranking given by another set of users from the same community. These recommendations are based on mostly a single-criterion rating or the average of overall ratings. These overall ratings avoid the implementation of fine grain analysis of the users’ behavior. The performance of any recommender system can be highly improvised by implementing the multi criteria recommender systems (MCRSs). In MCRSs, more feature reviews are taken into consideration in the algorithm making the recommendation process more efficient. Any comment given by the user online has a rich set of words that specify his preferences also known as review elements. There are many methods to extract these review elements, most widely used are Sentiment Analysis or Text Mining Methods. The recommender systems are more efficient if these review elements are incorporated in the algorithm. The major problems of RS like cold start problems and sparsity problems can be minimized substantially using this technique. The main aim of this paper is not only to understand these review items better, but explore methodologies of incorporating these elements into the RS to develop an efficient recommender system. It also explains how these elements can be incorporated in the recommender algorithms to enhance the RSs performance. This paper also presents the future trends to design efficient recommender systems using multi criteria decisions systems.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have