Clustering-based analysis of the relationship between clothing attributes and ease of styling

The aim of this study is to investigate the potential for diversifying recommendation results and improving user satisfaction by incorporating the concept of "ease of styling" into recommendation systems, and to further examine the relationship between "ease of styling" and the attribute information of clothing. Here, "ease of styling" refers to how easy it is to wear an item or an outfit well depending on its boldness. The results of the preliminary survey revealed that incorporating the concept of "ease of styling" into recommendations led to an increase in diversity and satisfaction under specific conditions. Furthermore, a clustering analysis was performed using clothing attribute information obtained from a survey related to "ease of styling" and an analysis of the characteristics of each cluster revealed specific combinations of clothing attributes related to "ease of styling". These findings are expected to serve as an important foundation for the future realization of a high-diversity clothing recommendation system that takes "ease of styling" into consideration.