首页|融合GWO和SVR的建筑安全事故预测模型

融合GWO和SVR的建筑安全事故预测模型

扫码查看
原文链接
  • NETL
  • NSTL
  • 万方数据
  • 维普
当前建筑业迅速发展，但随之而来的是频频发生的建筑安全事故，造成不可逆转的损失和伤害。虽然近些年来在建筑安全事故控制方面的研究已取得一定的成果，但建筑安全事故仍未得到有效控制。针对建筑业市政工程安全事故总数和死亡人数，探究二者之间的关系，构建灰狼优化算法-支持向量回归机(Grey Wolf Optimization and Support Vactor Regression，GWO-SVR)组合模型，收集 2008-2020年每个月的建筑安全事故数据及死亡人数数据集，发现二者之间成正向相关关系，以建筑安全事故数为特征对建筑死亡人数进行预测，精度达到95％以上，对建筑安全资源与人力投入有较大参考价值，有助于提升建筑安全管理水平。
Research on the prediction of integrating GWO-SVR methods model for construction safety accidents
Although many scholars have made considerable achievements in the safety management of construction accidents,there has not been any sound and appropriate tactic that fills the gap in the field of construction safety management.To explore the relationship between the total number of safety accidents and the number of deaths in public projects in the construction industry,an integrating prediction model based on Grey Wolf Optimization and Support Vactor Regression(GWO-SVR)is established.Firstly,the location of the grey wolf species is initialized,and the optimization parameters C and y are obtained by constant iteration on the finding of the optimal α,β,and δ.Next,the optimized parameters C and γ are replaced by the original parameters C and γ of the SVR model to produce the optimized prediction results.Finally,the predicted values of the GWO-SVR model is compared with the real values to determine the prediction accuracy.In this process,the optimized prediction results are compared with the predictions of the SVR model.A data set of construction safety accidents with the number of deaths in each month within 13 years from 2008 to 2020 is collected.There is a positive correlation between the two variables.The number of construction deaths is predicted using the number of construction safety accidents as a feature,and the prediction results of the model are evaluated using three evaluation indicators,SMEi,R2,and EMA i respectively.The validation results indicate that the developed model yields an accuracy of 95％in the prediction of construction deaths based on the characteristics of the construction safety accident data set.The accuracy is 0.3 higher than using a single SVR prediction model,and the indicators of SME,i and EMA.i are also better than those of the SVR model.The model has a better reference value for construction safety resources and manpower investment and helps to improve construction safety management.

safety social engineeringconstruction safety accidentsSupport Vactor Regression(SVR)Grey Wolf Optimization(GWO)model prediction

李政道、曾佳、吴恒钦

展开 >

深圳大学土木与交通工程学院,广东深圳 518060

安全社会工程 建筑安全事故 支持向量回归机 灰狼优化算法 模型预测

国家自然科学基金委员会项目广东省自然科学基金项目广东省教育厅项目

520783022021A15150122042021ZDZX1004

2024

10.13637/j.issn.1009-6094.2023.0503

安全与环境学报
北京理工大学 中国环境科学学会 中国职业安全健康协会

安全与环境学报

CSTPCD北大核心
影响因子：0.943
ISSN：1009-6094
年,卷(期)：2024.24(3)
  • 12