Abstract: This article proposes a knowledge graph-driven recommendation method that integrates user preferences and contrastive learning to enhance the performance of the recommendation system. The ...