基于大样本电梯抽查数据的风险评估Logistic模型

为了提高在用电梯监督抽查工作效率及有效性,在统计分析G市电梯安全监管抽查的大样本数据基础上,构建以电梯使用情况、电梯基本参数及制造维保相关情况等为指标的管理体系。根据电梯抽查数据的实质,先对数据进行变量筛选,然后构建风险分级,最后对前人的方法作出改进形成风险矩阵法并提出以Logistic方法为电梯整机风险建立量化模型,最终形成电梯整机风险评估体系。从理论的角度看,通过使用LIFT统计量和K-S统计量比较两种风险值计算模型,得出用Logistic方法进行风险分层更为准确。而实际的工程应用表明,利用Logistic回归法与基于平均风险值赋权比例法的组合为电梯安全监管抽样调查提供的筛选比例,比现有的方法更合理准确。

Logistic Model of Risk Assessment Based on Big Sampling Data of Elevator

In order to improve the efficiency and effectiveness of in-use elevator supervision and inspection,based on statistical analysis of the big sampling data of the G city,a management index system including elevator usage condition,basic parameters,maintenance,installation,and manufacturing is constructed in this paper.According to the essence of the elevator data,we selected variable firstly,then constructed risk classification criteria.At last we improve the previous method of the risk matrix and propose the Logistic method to establish quantitative models for the risk of elevator.Eventually a elevator risk assessment system is made.By comparing the two kinds of risk value calculation models with the LIFT statistic and the K-S statistic,it is concluded that the risk stratification is more accurate by using Logistic method.The results of application to real modeling show that the procedure we propose in this paper,by using the combination of the Logistic regression method and the average risk weighted ratio method for the elevator safety supervision sample survey,is more reasonable and accurate than the existing methods.