Evaluation Metrics for Personalized Recommendation Systems

Abstract

This paper summarizes and reviews the latest research progress on the evaluation metrics of the recommendation system, expounds from multiple perspectives such as accuracy, diversity, novelty, and conducts an in-depth analysis of their respective advantages and disadvantages and applicable environments, focuses on the three levels of recommended diversity and evaluation criteria, and makes predictions on the development direction of the evaluation metrics of the recommendation system.