생성형 AI를 이용한 서예 유튜브 콘텐츠 내 내재적 정보 추출 및 조회수에 미치는 영향에 관한 이질성 분석 연구

초록

This study aims to extract intrinsic content characteristics using generative AI reasoning beyond YouTube video external data collected via web crawling, and explore their effects on view counts. With NotebookLM, 42 external and intrinsic features were extracted from 237 calligraphy-related YouTube videos for impact analysis. The results showed two external and seven intrinsic factors significantly influenced view counts in the overall sample. Subgroup analysis indicated outstanding audience responses to large running script videos without background music, as well as long-duration large-font videos of scripts other than seal script created by male creators under the same condition. This study empirically proves generative AI can systematically extract intrinsic features of online video content, and provides practical implications for calligraphy creators to formulate content strategies for higher view counts.

키워드