• 목록

As population aging continues to intensify, home-based exercise has become an important way for older adults to maintain physical function and improve their quality of life. Virtual coaches can provide flexible exercise support; however, existing studies have mainly focused on functional aspects such as motion recognition and exercise feedback, while paying insufficient attention to users’ preferences regarding coach design. This study collected 25,684 comments from 15 older-adult exercise videos posted on WeChat official accounts. After text preprocessing, 13,427 valid comments were analyzed using text mining and Latent Dirichlet Allocation (LDA) topic modeling. The results identified five major user requirements: cultural and emotional atmosphere, movement demonstration and safety guidance, aesthetic expression and professionalism, pace adjustment and patient instruction, and professional identity and demeanor. Based on these findings, this study proposes design directions for virtual coaches in terms of character image, movement presentation, instructional style, and interaction design. The findings provide an evidence-based foundation for designing home-based exercise support systems for older adults.