基于RBF神经网络的震伤人员快速评估模型

发布时间：2018-05-02 07:48

本文选题：震灾 + 震伤人员评估　；参考：《系统工程》2016年08期

【摘要】：震伤人员的快速评估对统筹配备医疗资源和进一步预测废墟受伤和失踪人数十分关键。为提高震后受伤人员评估结果的可靠性,通过对影响震伤的关键性影响因子的提取,采用能够有效处理模糊性和非线性指标的RBF神经网络模型对震后受伤人员进行快速评估。从分析震后造成人员受伤的影响因子入手,从承载体减抗风险能力、暴露性和敏感性三个维度提出震伤人员预测指标体系;在受伤人员预测方法上,考虑指标的小样本性、非线性和部分指标的模糊性特征,将模糊逻辑与神经网络方法结合起来,采用动态优化的径向基(RBF)神经网络方法,以提高评估模型的全局搜索和优化能力,避免常规BP神经网络较早陷入局部优化的不足;案例结果显示:与BP神经网络训练的绝对误差3.24%相比,RBF神经网络震伤人员评估模型的绝对误差能够降低至1.71%,精度提高47.2%,说明本文评估模型评估可靠性高、模型鲁棒性强,能够推广于震灾应急的管理决策之中。
[Abstract]:Rapid assessment of earthquake victims is critical to co-ordinating medical resources and further predicting the number of debris injured and missing. In order to improve the reliability of the evaluation results of post-earthquake injured people, the RBF neural network model, which can deal with the fuzzy and nonlinear indexes effectively, was used to evaluate the injured persons quickly by extracting the key influencing factors. Based on the analysis of the influencing factors of human injury after the earthquake, the prediction index system of earthquake injury personnel is put forward from the three dimensions of risk reduction, exposure and sensitivity of the carrier, and the small sample nature of the index is considered in the prediction method of the injured person. In order to improve the global searching and optimization ability of the evaluation model, the nonlinear and fuzzy characteristics of some indexes are combined with fuzzy logic and neural network method, and the radial basis function (RBF) neural network method of dynamic optimization is adopted to improve the global searching and optimization ability of the evaluation model. To avoid the deficiency of conventional BP neural network falling into local optimization earlier; The results show that compared with the absolute error of BP neural network training 3.24%, the absolute error of earthquake personnel evaluation model of RBF neural network can be reduced to 1.71 and the accuracy is increased by 47.2%. It shows that the evaluation model in this paper has high reliability and strong robustness. It can be extended to the management decision of earthquake emergency.
【作者单位】：西南科技大学经济管理学院;东北林业大学土木工程学院;
【基金】：教育部人文社科研究基金资助项目(16YJC630040) 省教育厅人文社科重点资助项目(15SA0034)
【分类号】：P315.9;TP183

【相似文献】