A Convergence Study on User Preference Topics for 3D Animated Virtual Humans on Short Video Platforms - An LDA Text Mining Analysis Based on Liu Ye Xi -

다국어 초록 (Multilingual Abstract)

3D animated virtual humans, as an innovative application of artificial intelligence, 3D modeling, and animation technology, are leading new trends in digital entertainment and social interaction. On short video platforms, these virtual characters, which combine anime art with 3D technology, attract substantial user attention through detailed modeling design, smooth animation performance, and engaging emotional interactions. However, user preferences for 3D animated virtual humans and their evolution patterns have not been systematically studied. This research focuses on Liu Ye Xi, a 3D animated virtual human on the Douyin platform, using LDA Topic Model to conduct text mining analysis on 52,387 user comments. The research reveals that user preferences exhibit multi-dimensional characteristics, primarily concentrated in eight thematic dimensions: Virtual Character's Emotional Connection (24%), User's Interactive Experience (19%), Digital Life's Humorous Expression (16%), Cultural Phenomenon on Social Platforms (12%), Integration of Technology and Innovation (10%), Real-World Impact of Digital Virtual Humans (8%), Growth and Self-Identity (7%), and Emotional and Psychological Resonance (4%). These preferences show an evolutionary trend from surface attraction to deep thinking to emotional resonance over time.
This finding not only enriches the theoretical framework of 3D animated virtual humans research and reveals the dynamic evolutionary characteristics of user preferences but also provides empirical evidence for the design development and content creation of 3D animated virtual characters. These research results have important implications for understanding the future development direction of virtual entertainment and the application of 3D animated virtual humans in social media.