LDA토픽모델링을 적용한 성인발레 콘텐츠 동향분석: 2018~2025 유튜브 데이터를 중심으로

초록

This study examined YouTube discourse on ‘adult ballet’ from 2018 to 2025 (8 years) by applying Latent Dirichlet Allocation(LDA) topic modeling, on the top 100 videos per year. Date collected NetMiner v4 were preprocessed and analyzed with the u_mass coherence metric and Spearman’s rank-order correlation. The analysis revealed four ma jor topics: (1) learning & performance (practice, skill acquisition, achievement); (2) balletwear, equipment & con sumer culture (leotards, gear, recommendations); (3) physical effects & healthy (alignment, diet/weight control); (4) everyday practice & digital diffusion (vlogs, challenges, tips). Only Topic 2 showed a significant positive correlation with time(ρ = 0.347, p = 0.042), indicating a growing prominence of consumption and equipment focused discourse. Consequently, the adult ballet content has been reorganized toward greater emphasis on consumption and participation. To build a reliable digital arts learning culture, educators, institutions, and creators should stand ardize informaion such as sizing guides and care tips and ensure transparency in sponsorship disclosure.

키워드