Deep Content-Collaborative Recommender System (DCCRS)

Abstract

With the ongoing growth, involution and dynamicity of online infrastructure, recommender systems have been an efficacious key solution for better personalization. Most online services have begun strategizing how to provide users with more personalized recommendations. Recommendation quality and system scalability are addressed in most recommender systems by two popular approaches, Collaborative Filtering and Content-based methods. Both methods have their own rewards, while in isolation; they fail to provide quality recommendations in some situations. Therefore, the aim of proposed approach is to put forward an elegant and effective framework that combines components from both the methods in cross domain framework. Alongside, the paper infuses the power of Deep Learning into recommender system to learn interesting mappings of users and items in a shared semantic space. This fresh approach is termed as Deep Content-Collaborative Recommender System (DCCRS) over cross domain.