基于随机森林的非常规突发事件结构化情景体系模型  

Structured Scenario System Model for Unconventional Emergency Based on Random Forest

作　　者：张咏华 邬开俊[1] ZHANG Yong-hua;WU Kai-jun(School of Electronics and Information Engineering,Lanzhou Jiaotong University,Lanzhou,Gansu 730070,China)

机构地区：[1]兰州交通大学电子与信息工程学院,甘肃兰州730070

出　　处：《公路交通科技》2024年第2期139-147,190,共10页Journal of Highway and Transportation Research and Development

基　　金：国家社会科学基金项目(14CGL001);甘肃省委组织部人才项目(2023-02)。

摘　　要：为降低非常规突发事件发生所造成的各类损失,构建基于随机森林的非常规突发事件结构化情景体系模型。根据随机森林相关理论知识,使用python进行了试验仿真,并将交通隧道工程的坍塌事故视为非常规突发事件进行处理。首先,针对传统的非常规突发事件结构化情景模型存在的问题,构建非常规突发事件结构化描述空间,利用随机森林模型对非常规突发事件进行了分级,按照分级的结果,进行非常规突发事件的结构化描述。其次,基于朴素贝叶斯模型与支持向量机模型进行了对比试验,在相同的试验环境下,使用不同的数据集通过多次试验分析比较。最后,基于随机森林的回归模型预测事故类型,使用决策树模型对数据进行了可视化分析,并计算使用重要特征预测事故类型的时间效率与准确率。利用实际发生的交通隧道工程坍塌事故验证了随机森林回归模型的实用性。结果表明:随机森林模型的平均准确率比其他两种模型提高了至少8.77%;仿真结果与真实值几乎相近,仿真准确率值为1。研究结果为国家应急管理部门提供了应急决策依据,帮助决策者在最短的时间内提出合理的应急响应和应急预警方案,可以有效减少非常规突发事件发生所造成的经济损失,因此随机森林模型有较高的使用价值。In order to reduce the losses caused by the occurrence of unconventional emergencies,a structured scenario system model of unconventional emergencies based on random forest is constructed.According to the theoretical knowledge of random forest,python is used to conduct experimental simulation,and the collapse accident of traffic tunnel engineering is treated as an unconventional emergency.First of all,aiming at the problems of the traditional structural scenario model of unconventional emergencies,the structured description space of unconventional emergencies is constructed,the random forest model is used to classify the unconventional emergencies,and the structured description of the unconventional emergencies is carried out according to the classification results.Secondly,a comparative experiment based on the Naive Bayes model and the support vector machine model is constructed,in the same experimental environment,different data sets are used for multiple experimental analysis and comparison.Finally,the accident type is predicted based on the random forest regression model,the decision tree model is used to visualize the data,and the time efficiency and accuracy of predicting accident type using important features are calculated.The practicability of the random forest regression model is verified by the actual traffic tunnel collapse accident.The result shows that(1)the average accuracy of the random forest model is improved by 8.77% compared with the other two models;(2)the simulation results are close to the real values,and the simulation accuracy value is 1.The study result provides emergency decision-making basis for national emergency management departments,and helps decision-makers to put forward reasonable emergency response and emergency earlywarning plans in the shortest time,it can effectively reduce the economic losses caused by non-conventional emergencies,therefore,the random forest model has high value.

关 键 词：隧道工程 结构化情景体系 随机森林 非常规突发事件 事件类型 

分 类 号：U458.3[建筑科学—桥梁与隧道工程]