PROJECT TITLE :

A Survey on Knowledge Graph-Based Recommender Systems

ABSTRACT:

The issue of an excessive amount of information has prompted the development of recommender systems, which model users' preferences in order to provide an improved user experience across a variety of online applications. Although a great number of efforts have been made in the direction of providing more personalized recommendations, recommender systems continue to struggle with a number of difficulties, including data sparsity and cold-start problems. In recent years, there has been a significant uptick in interest regarding the generation of recommendations using the knowledge graph as supplementary information. Not only can this strategy help alleviate the problems that were mentioned above, which will result in a more accurate recommendation, but it can also provide explanations for the items that were recommended. In this paper, we conduct a comprehensive analysis of different types of recommender systems that are based on knowledge graphs. We compile recently published papers in this area and organize them into three distinct categories: connection-based methods, embedding-based methods, and propagation-based methods. In addition to this, we further subdivide each category based on the characteristics of these different methods. In addition, we investigate the proposed algorithms by concentrating on how the papers make use of the knowledge graph in order to make recommendations that are accurate and easy to explain. In conclusion, we present a number of potential lines of inquiry in relation to this field.