Incomplete preference relations to smooth out the cold-start in collaborative Recommender Systems

E-commerce companies have developed tools to assist users in finding the most suitable items for their needs or preferences. The most successful tool in this area has been the recommender systems. This kind of software obtains information about the users´ tastes, opinions, necessities, and with a recommendation algorithm, infers recommendations that lead users to the most suitable items for them. These algorithms usually require a significant quantity of information and that information is not always available or easy to obtain. Overcome this problem, known as the cold-start problem, and reduce the requirements of information is not an easy task. In this contribution, we review this topic and present our proposal: a hybrid recommender system which combines a collaborative filtering algorithm with a knowledge-based one in order to improve the cold-start problem.