A Critical Analysis of Collaborative Topic Regression Based Recommender Systems

Abstract

Today the recommendation technology has become an integral part of modern digital world. Its impeccable role is clearly visible in online applications ranging from advertisement to e-commerce applications, matrimonial sites to social networking etc. This technology employs different recommendation methods to recommend items, products or services to the target user. Among them, Collaborative Filtering is the only technique that has become the first choice of researchers and has managed to secure a unique place in the recommendation systems. It mainly relies on user-item feedback matrix to make recommendations which is most of the time found extremely sparse and hence causes the sparsity problem. Thus, to address this issue, a lot many researches have been carried out based on Collaborative Topic Regression (CTR) in recent times. These researches have achieved encouraging performance by fusing a variety of auxiliary information into the modeling framework. The present study aims to provide the insights and critically analyze the recent trends in CTR based recommendation systems by investigating the published lit­erature.