基于GA-SVM的施工人员安全行为影响因素及决策模型研究

发布时间：2018-03-13 01:36

本文选题：安全行为　切入点：影响因素　出处：《中国安全生产科学技术》2014年12期 　论文类型：期刊论文

【摘要】：导致施工人员不安全行为的因素众多,如何保证员工进行安全施工是施工企业亟待解决的问题。为分析施工人员安全行为的影响因素及作用机理,从社会资本理论、认知心理学理论和安全行为理论分析了安全行为的影响因素,并设计了相关调查问卷。在进行仿真分析前,采用遗传算法优化计算的方法筛选出了13个关键影响因素,降低了自变量之间的相关性,之后利用支持向量机(SVM)的方法对决策模型进行了仿真分析,并与BP神经网络模型做了对比。仿真结果表明:基于筛选出的关键影响因素的SVM仿真模型的精度和有效性大于BP神经网络模型,模型精度为0.00887,相关系数为88.4%,说明影响因素与安全行为之间具有较好的拟合关系。研究结论为企业衡量员工安全行为水平,提高员工安全行为能力和企业安全管理能力提供理论支持。
[Abstract]:There are many factors leading to the unsafe behavior of construction personnel. How to ensure the safety of construction workers is an urgent problem to be solved. In order to analyze the influencing factors and the mechanism of safety behavior of construction personnel, the social capital theory is introduced in this paper. The cognitive psychology theory and the safety behavior theory analyze the influencing factors of safety behavior, and design the related questionnaire. Before the simulation analysis, 13 key influencing factors are selected by genetic algorithm optimization method. The correlation between independent variables is reduced, and then the decision model is simulated by using support vector machine (SVM) method. The simulation results show that the SVM simulation model based on the selected key factors is more accurate and effective than the BP neural network model. The precision of the model is 0.00887, and the correlation coefficient is 88.4, which indicates that there is a good fitting relationship between the influencing factors and the safety behavior. Provide theoretical support to improve employee safety behavior ability and enterprise safety management ability.
【作者单位】：天津财经大学商学院;天津财经大学经济学院;
【基金】：国家自然科学基金项目(71171140) 天津市科委项目(12JCZDJC34900) 天津财经大学重点科研项目(ZD1306) 天津财经大学研究生科研资助计划项目(2014TCB05)
【分类号】：TU714

【相似文献】