Data-driven classification of lower body shapes in older women

This study uses advanced data science and machine learning techniques to explore the lower body morphology of older women and provide a basis for personalized clothing design. Using the VITUS body scanning system and contact measurement method, data on 26 lower body characteristic variables were collected from 217 older women. Box plots and Q-Q plots in SPSS were used to screen outliers and test for normal distribution. The principal component analysis method was used to identify five key characteristic parameters for lower body shape classification: abdominal height, hip circumference, hip-waist difference, lateral hip angle, and Upper hip convexity angle. A two-step clustering algorithm divided the lower body shape into three categories based on the selected characteristic parameters. Finally, the discriminant analysis was used to verify the model, and the test results showed that the highest accuracy was 93.3%. The results show that the classification rules and methods can lay the foundation for the subsequent older women's clothing design system.