Short Video Account Influence Evaluation Model Based on Improved SF-UIR Algorithm

In view of the lack of numerical value of the influence of the account’s own behavior and the lack of the influence of following fans based on the topological structure platform in the evaluation of the influence of short video accounts, this paper proposes a short video platform account influence calculation algorithm FF-SF-UIR (Factor analysis and Fan group chat in SF-UIR) based on the improved SF-UIR (Self and Followers User Influence Rank) algorithm from three perspectives: the influence of the official platform label authentication, the influence of the work, and the influence of the following fans based on the topological structure platform. This method analyzes the rationality of the historicity and periodicity of the influence of the work by using the numerical influence of the official label of the account platform and introducing the factor analysis method. According to the social network topology, the influence of followers based on the topological structure platform is introduced, which makes the evaluation result more reasonable and provides a new way for influence ranking of short video accounts. Taking Douyin (TikTok in China) short video platform as the experimental object, the experimental results show that compared with other existing methods, this method improves the precision by 9.67%, the recall by 23.67%, and the F1-Score by 20.68% in the evaluation index.