Prediction of New Media Communication of Fitness Culture and Its Influence Factors

Abstract

As Bilibili became influential, its online users have gradually developed their fitness habits. However, the content distribution mechanism for fitness videos on new media platforms has not been well developed, resulting in many excellent works being overwhelmed by vast amounts of videos. To promote fitness culture to be better developed and spread, this study analyzes 973 fitness videos on Bilibili, explores the factors influencing fitness video communication, and constructs a predictive model. Firstly, based on the Elaboration Likelihood Model (ELM), this paper builds a theoretical model of the influence factors around the central and peripheral routes. Secondly, the XGBoost algorithm is used to construct the predictive model of communication effectiveness of fitness videos and determine the important influence factors according to the feature importance. Thirdly, we use regression analysis to explore how the important factors influence communication effectiveness. The results show that videos with a weight loss theme can improve the overall communication effect, while life-sharing videos will reduce the effect. Subtitles negatively affect the overall communication effect, while video duration and fan base can significantly improve the overall communication effect. Female fitness uploaders’ videos are less effective in spreading, and Chinese fitness uploaders’ videos are more effective.