支持向量机的建筑施工安全预警模型仿真

Simulation of Building Construction Safety Early Warning Model Based on Support Vector Machine

针对传统的建筑施工安全预警模型存在响应时间较长、误差较大等问题,构建了一种基于支持向量机(SVM)的建筑施工安全预警模型。分析所研究建筑工地中存在的危险源,以危险源以及安全防护措施作为切入点,针对不同的建筑安全预测对象的不同特征,构建新的建筑工地预警指标体系。对预警指标体系中的安全因素进行预处理,将经过处理后的信息结构作为SVM的输入数据进行训练以及预测,构建建筑施工安全预警模型。实验结果表明,所提模型的预警时间在20min以下,误差较低,能够达到期望的预警效果。

Due to long response time and large error in the traditional model, this article proposed an early warning model for construction safety based on support vector machine ( SVM). Firstly, the dangerous source in the researched construction site was analyzed. Secondly, dangerous source and safety protection measure were used as the points of penetration. Thirdly, new early warning indicator system of construction site based on different characteristics of different building safety prediction objects. Moreover, safety factors in the early warning indicator system were preprocessed. And then, the processed information structure was trained and predicted as the input data of SVM. Thus, the early warning model of building construction safety was built. Simulation results show that the early warning time of the proposed model is less than 20 minutes. This model has low error, which can achieve the desired early warning effect.