Multi-objective optimization with recommender systems: A systematic review

Abstract

Recommender systems have become essential in modern information systems and Internet applications by delivering personalized and pertinent content to users. While conventional recommendation algorithms usually prioritize optimizing a single objective, it is now evident that considering additional metrics is crucial for improving the overall user experience. Despite the importance of considering multiple objectives, conventional recommendation models face the challenge of balancing these objectives, which can sometimes conflict with each other. To tackle this challenge, there is a growing interest in multi-objective recommender systems (MORS) that consider multiple objectives simultaneously and provide a more personalized and varied set of recommendations. MORS can optimize recommendations based on various metrics, including accuracy, diversity, novelty, and user satisfaction, leading to more efficient and personalized recommendation systems. The objective of this paper is to conduct a systematic review study to assess the current state of research in the field of MORS and identify potential avenues for future exploration. The study selection procedure includes 78 primary studies published from 2019 to January 2023. These preliminary studies are categorized based on different variables to address the research questions outlined in this study. The findings of this systematic review study reveal a diverse range of applications, objectives, datasets, methodologies, and evaluation metrics utilized in the field of MORS. Additionally, this review offers a crucial overview of the current state of research in this area, highlighting the existing challenges and future directions for enhancing the efficiency of MORS. These outcomes can benefit both professionals and academic researchers in the development and implementation of effective MORS.