@article{Deng2026, 
author = {Zixuan Deng and Mingqi Zhuang and Zhanming Zhang},
title = {Research on Short-Term Risk Prediction for Temporary Performance Scenes Based on PSO-SVM},
year = {2026},
journal = {Journal of Xinjiang University(Natural Science Edition in Chinese and English)},
volume = {43},
number = {5},
pages = {577-587},
keywords = {temporary performance scenes, particle swarm optimization, support vector machine, risk prediction},
url = {https://www.sciopen.com/article/10.13568/j.cnki.651094.651316.2026.05.06.0001},
doi = {10.13568/j.cnki.651094.651316.2026.05.06.0001},
abstract = {Temporary performance scenes, owing to their unfixed structures, variable environments, and rapid erection, often face higher safety risks than permanent buildings, and are prone to problems such as structural tilting and settlement. To address those problems, this paper proposes a short-term risk prediction method for temporary performance scenes based on optimizing support vector machine (SVM) using particle swarm optimization (PSO). Firstly, introduces the basic principle of SVM, determines the variables to be optimized, and provides the optimization principle of PSO algorithm; Then, designs a temporary performance scene risk prediction model based on PSO-SVM. Selects key state information such as structural inclination, settlement, and electrical equipment current/temperature, etc., builds a data acquisition experimental platform; Finally, uses the established risk prediction model to analyze, train, and predict the collected data, and conducts experimental data validation. The results show that the PSO-SVM method not only achieves high prediction accuracy in the training set, validation set, and prediction set (all above 96%), but also enhances its practicality and decision credibility in engineering risk monitoring through the fusion of confidence intervals and threshold warning mechanisms.}
}