Development of an Optimal Mattress Recommendation Model Considering User Body Shape and Mattress Experience

This study aimed to develop a recommendation model for selecting the optimal type of mattress by simultaneously considering users' body shapes and mattress usage experiences, thereby contributing to improved sleep quality. An experiment was conducted with 176 adults who had more than five years of mattress usage experience. The experiment was designed to derive priority rankings for eight different mattress types based on user body characteristics and mattress usage experiences. Machine learning algorithms including Decision Tree, Random Forest, XGBoost, and LightGBM were employed to build the recommendation model, and its performance was evaluated using the NDCG. The results showed that the XGBoost model achieved the highest performance (NDCG@1 = 0.7139) after relabeling. Although this performance is not exceptionally high in absolute terms compared to previous similar studies, it demonstrates meaningful efficacy as a personalized recommendation model in practical mattress selection scenarios. These findings highlight the importance of an integrated approach combining physical and experiential factors and are expected to serve as foundational data for developing personalized recommendation systems in the mattress industry.