中国安全科学学报 ›› 2020, Vol. 30 ›› Issue (4): 114-120.doi: 10.16265/j.cnki.issn1003-3033.2020.04.018

• 公共安全 • 上一篇    下一篇

无序多分类Logistic回归的地铁扶梯事故特征分析

王志如1 讲师, 高林源1, 王敏2 高级工程师   

  1. 1.北京科技大学 经济管理学院,北京 100083;
    2.北京市地铁运营有限公司,北京 100044
  • 收稿日期:2020-01-05 修回日期:2020-03-12 出版日期:2020-04-28 发布日期:2021-01-27
  • 作者简介:王志如(1984—),女,内蒙古鄂尔多斯人,博士,讲师,主要从事地铁安全、地铁网络脆弱性、应急管理等方面的研究。E-mail:wangzhiru@ustb.edu.cn。
  • 基金资助:
    国家自然科学基金资助(51978164);北京市自然科学基金资助(L181009,9194028);教育部人文社科基金资助(18YJC630193)。

Characteristic analysis of subway escalator accidents based on disordered multinomial Logistic regression

WANG Zhiru1, GAO Linyuan1, WANG Min2   

  1. 1. School of Economics and Management, University of Science and Technology Beijing, Beijing 100083, China;
    2. Beijing Subway Co., Ltd., Beijing 100044, China
  • Received:2020-01-05 Revised:2020-03-12 Online:2020-04-28 Published:2021-01-27

摘要: 为探究影响地铁扶梯事故的主要因素,以北京地铁扶梯事故数据为基础,利用无序多分类Logistic回归方法,研究地铁扶梯事故的特征,通过控制相关因素来预防地铁扶梯事故的发生;分析894起地铁扶梯事故的事故形式、解释变量与响应变量的类型及属性,构建无序多分类Logistic回归模型,识别主要影响因素,计算各类影响因素对事故的贡献大小。研究结果表明:针对不同类型事故,制定不同的环境因素、乘客特征、乘客行为因素、线路因素的重点管控策略,可降低地铁扶梯事故的发生;该研究方法能够用于地铁扶梯事故预测研究,解决变量无序且多元的相关性分析和贡献大小的定量分析问题。

关键词: 地铁扶梯事故, 无序多分类, Logistic回归分析, 事故特征, 相关性分析, 解释变量

Abstract: In order to explore major reasons for subway escalator accidents, with data of such accidents in Beijing subway as an example, disordered multinomial Logistic regression method was used to study characteristics of accidents and to prevent their occurrence by controlling related influencing factors. Then, by analyzing accident forms, types and attributes of explanatory and response variables in 894 subway escalator accidents, a disordered multinomial Logistic regression model was constructed to identify significant correlation factors and calculate their contribution level to accidents. The results show that developing different key control strategies concerning environmental factors, passenger characteristics, passenger behaviors and paths for different types of accidents can effectively reduce occurrence of subway escalator accidents. At the same time, this method can be applied in research of accident predication as well as solve problems of correlation analysis of disordered multivariant variables and quantitative analysis of their contribution.

Key words: subway, escalator accident, disordered multidassification, Logistic regression analysis, accident characteristic, correlation analysis, explanatory variables

中图分类号: