基于PCA-Logit模型的无车承运人风险预警  

Risk warning of no-car operating carrier based on PCA-Logit model

作　　者：常连玉[1,2] 胡大伟[1] 陈海燕[2] 孔月红[3] 

机构地区：[1]长安大学汽车学院,陕西西安710064 [2]交通运输部管理干部学院,北京065201 [3]辽宁省交通高等专科学校,辽宁沈阳110122

出　　处：《江苏大学学报（自然科学版）》2017年第3期273-279,共7页Journal of Jiangsu University：Natural Science Edition

基　　金：国家自然科学基金资助项目(51278057);交通运输部应用基础研究项目(2014319226270);辽宁省交通高等专科学校优秀人才支持项目(2015lnccrcky06)

摘　　要：为了评估无车承运人承运经营的风险,提高风险预警的准确性,提出了PCA-Logit无车承运人风险预警模型.根据无车承运人经营过程中风险的主要来源,构建了包含26项指标的风险预警指标体系,在对指标体系进行主成分分析(PCA)的基础上,通过Logit风险预警模型计算无车承运人的风险概率.选取15家无车承运人企业进行案例分析,提取了累计方差贡献率达91.03%的8个主成分,根据因子载荷矩阵值得出8个主成分的表达式,采用最大似然估计法对Logit风险预警模型进行参数标定,并计算得到15家无车承运人企业的风险预警值.结果表明:无车承运人企业的风险概率与企业的管理水平、运输市场的市场竞争强度、托运人及实际承运人谈判价格的优势和托运货物的属性等因素具有较强的相关性.To estimate the risk and improve the accuracy of risk warning for no-car operating carrier, the PCA-Logit risk warning model was established. The risk warning index system with 26 indices was put forward according to the main source of risk in no-car operating carrier business prorisk warning index system, and the risk probability was calculated by Logit risk warningoperating carrier enterprises were chosen for case study , in which 8 principal components with cumulative variance contribution rate of 91.0 3 % were extracted to obtain the expressions factor loading matrix. The model parameters were calibrated by the maximum likelihood estimation, andthe risk warning value of 15 no-car operating carrier enterprises were calculated. The results show thatthere is strong correlation among no-car operating carrier risk and management level, competition strength of transportation market, negotiating price advantage of shipper and actual carrier and attribute of goods.

关 键 词：无车承运人 道路运输 风险预警 PCA-Logit模型 最大似然估计 

分 类 号：U491[交通运输工程—交通运输规划与管理]