A Survey on Knowledge Graph-Based Recommender Systems

Recommender systems have emerged as indispensable tools for information filtering, and the integration of knowledge graphs for auxiliary information is becoming an increasingly popular research topic. This paper reviews recent studies, discussing the current state and practical applications of knowledge graph-based recommender systems. We summarize the strengths and weaknesses of various knowledge graph-based recommendation methods, noting that these systems significantly enhance performance in areas like accuracy, diversity, interpretability, and novelty. Furthermore, the trend of combining different knowledge graph-based methods underscores the mainstream evolution of recommender systems, warranting future exploration. We finish with an analysis of current challenges and a forward-looking perspective on future advances. This review aims to assist the reader in understanding and navigating this research field.

I. Introduction

The structure of this paper is as follows: Section I introduces the background of the recommendation system based on knowledge graph. Section II introduces related knowledge from two aspects of recommendation systems and knowledge graph-based recommender systems. Section III and Section IV classify and sort out the recommendation methods based on knowledge graph in recent years, make tables for reading and searching, discuss and analyze them respectively, summarize the advantages and disadvantages of various methods, and predict the development trend in this field. Section V introduces the application of knowledge graph-based recommendation methods in various fields. Section VI discusses the existing problems and development prospects of recommendation methods based on knowledge graph. Finally, the paper is concluded in Section VII.