Outline

Ingegneria Sismica

Ingegneria Sismica

Propagation Mechanism and user behavior analysis of short video content in mobile social platform

Author(s): Wei Zhu1, Gui Wang2, Xuan Hao1, Munire Adili3, Weiqun Cui1
1Department of Information Engineering, Shandong Water Conservancy Vocational College, Rizhao 276800, Shandong, China
2Office of Registrar, Zibo Polytechnic University, Zibo 255300, Shandong, China
3School of Digital Information Engineering, Kashgar Vocational and Technical College, Kashgar 844000, Xinjiang, China
Zhu, Wei. et al “Propagation Mechanism and user behavior analysis of short video content in mobile social platform.” Ingegneria Sismica Volume 43 Issue 1: 1-22, doi:10.65102/is2026315.

Abstract

Aiming at the problem of the separation analysis of short video propagation process and user behavior response in mobile social platforms, a comprehensive analysis framework combining content characteristics, platform mechanism, social structure and behavior feedback was constructed. Based on 52,000 short video samples and corresponding user behavior logs collected from January 2024 to December 2024, multi-modal content representation, propagation network analysis, clustering identification and predictive modeling methods are used to jointly test the propagation trigger mechanism, diffusion evolution mechanism and user behavior transformation path. The results show that topic focus, emotion expression intensity, information density and multi-modal presentation quality form the trigger basis of propagation, and the recommendation strength, hot traffic allocation and social relationship strength significantly affect the speed, scope and persistence of diffusion. Users show obvious heterogeneity in the dimensions of click, stay, comment, forward, follow and repeat visit. Low-active users are more dependent on the accuracy of recommendation, and high-active users are more likely to be stimulated by social interaction. Experimental results show that the Accuracy, F1 and AUC of the constructed model on the user behavior recognition task reach 0.903, 0.895 and 0.941, respectively, and the RMSE and R2 on the propagation effect prediction task are 0.109 and 0.872, respectively. This study can provide quantitative basis for short video propagation mechanism explanation, user behavior modeling and platform content governance.

Keywords
Short video propagation; User behavior; Recommendation algorithm; Multimodal analysis

Related Articles

Huiqiao Liu1
1Yinchuan University of Energy, Ningxia, 750000, China
Xin Zhao1, Yan Li1, Xiangyang Cao1, Qiushuang Li1, Jianing Zhang1
1State Grid Shandong Electric Power Company Economic and Technological Research Institute ShanDong JiNan 250001, China
Dan Yang1
1School of Marxism, Suzhou Polytechnic University, Suzhou, 215104, China
Liuhang Shen1, Xiangwen Sun1
1Ulster college at Shaanxi University of Science &Technology, Xi’an,710021, Shaanxi, China