Cross Domain Recommendation Systems using Deep Learning: A Systematic Literature Review

Abstract

Recommender Systems play an inevitable role nowadays as in every field, there are plethora of choices to consider. With the advent of more and more Artificial Intelligence and Machine Learning practices, recommender systems are rising as a blooming area with vast horizon. Cross Domain Recommender Systems are a way to learn user’s interests and choices from source domain(s), transfer the learnt features to a target domain from which recommendations are to be made. Cross Domain Recommender Systems not only make effective recommendations, but also solve Cold Start and Data Sparsity problems. Deep learning based recommender systems are able to make recommendations from large datasets and are in trend. The presented work fulfils the need of a systematic survey of all the Cross Domain Recommender System implemented using Deep Learning Techniques. After excluding all the papers based on only Cross Domain Recommender Systems (without Deep Learning), Deep Recommender Systems (working within one domain only) and all other classical Recommender Systems from the research papers available over all the quality sources from the web, 42 papers were shortlisted and studied for the techniques being used, domains and datasets on which CDRS using Deep Learning have been implemented, research issues and the evaluation parameters on which Deep Learning based CDRS can be verified.