基于零膨胀回归模型的城市轨道交通触网故障频次研究
|
摘要
城市轨道交通触网故障后果严重,但发生频次较少,数据分析困难.零膨胀计数模型(ZIM)对零值大的数据集具有良好的适用性.针对上海地铁近4年运营过程中积累的触网故障数据进行统计分析,采用ZIM模型中运用最广泛的ZIP模型和ZINB模型进行建模,对比模型的4项评价指标,并进行模型命中率、泛化能力、释义合理性的评价.研究表明,ZINB模型能够对触网故障数据进行更好的拟合.基于模型结果,对城市轨道交通触网系统的安全运营策略及维修保养制度提出建议.
The consequences of the fault in the contact line system of urban rail transit are serious, but the frequency of occurrence is low, so it is difficult to analyze the fault. The zeroexpansion count model(ZIM) has good applicability to zero-valued data sets. We analyzed statistical data on contact line system fault accumulated during the recent four-year operation of Shanghai Metro, tried the most widely used ZIP model and ZINB model, compared the four evaluation indicators, and evaluated model hit rate, generalization ability, interpretation rationality separately. The research shows that the ZINB model can better fit the contact line system fault data. Based on the results of the model, suggestions are given for the safety operation strategy and maintenance system of the urban rail transit contact line system.
The consequences of the fault in the contact line system of urban rail transit are serious, but the frequency of occurrence is low, so it is difficult to analyze the fault. The zeroexpansion count model(ZIM) has good applicability to zero-valued data sets. We analyzed statistical data on contact line system fault accumulated during the recent four-year operation of Shanghai Metro, tried the most widely used ZIP model and ZINB model, compared the four evaluation indicators, and evaluated model hit rate, generalization ability, interpretation rationality separately. The research shows that the ZINB model can better fit the contact line system fault data. Based on the results of the model, suggestions are given for the safety operation strategy and maintenance system of the urban rail transit contact line system.
引文
[1]沈丽莉.地铁供电系统可靠性评估与安全评价方法的研究[D].大连:大连交通大学,2013.
[2]曾平,刘桂芬,曹红艳.零膨胀模型在心肌缺血节段数影响因素研究中的应用[J].中国卫生统计,2008,25(5):464-466.
[3]徐建,孙璐.解决交通事故数据分析中零值问题的模型[J].吉林大学学报,2015, 45(3):769-775.
[4]项丽雅.ZIP回归模型的变量选择及其在保险匣定中的应用[D].上海:复旦大学,2012.
[5]缪柏其,韦剑,宋卫国,等.林火数据的Logistic和零膨胀Poisson(ZIP)回归模型[J].火灾科学,2008,17(3):143-149.
[6]徐涛,李卫,陈涛.零频数过多模型在亚健康状态研究中的应用[J].中华流行病学杂志,2011,32(2):187-191.
[7] Daniel B.Hall.Zero-Inflated Poisson and Binomial Regression with Random Effects:A Case Study[J].Biometrics, 2000, 56(4):1030-1039.
[8]王国梁.高速铁路特殊咽喉区接触网布置研究[J].铁道工程学报,2017(10):73-77.
[9]杨桢.接触网分段绝缘器的动态特性分析及维护[J].电气化铁道,2011(2):1-4.
[10]沈修建,吕继涛.刚性悬挂接触网设计若干技术问题探讨[J].电气化铁道,2004(3):60-62.
[11]李雪昆.故障预测与健康管理技术在地铁列车上的应用[J].城市轨道交通研究,2018(2):105-107.
[2]曾平,刘桂芬,曹红艳.零膨胀模型在心肌缺血节段数影响因素研究中的应用[J].中国卫生统计,2008,25(5):464-466.
[3]徐建,孙璐.解决交通事故数据分析中零值问题的模型[J].吉林大学学报,2015, 45(3):769-775.
[4]项丽雅.ZIP回归模型的变量选择及其在保险匣定中的应用[D].上海:复旦大学,2012.
[5]缪柏其,韦剑,宋卫国,等.林火数据的Logistic和零膨胀Poisson(ZIP)回归模型[J].火灾科学,2008,17(3):143-149.
[6]徐涛,李卫,陈涛.零频数过多模型在亚健康状态研究中的应用[J].中华流行病学杂志,2011,32(2):187-191.
[7] Daniel B.Hall.Zero-Inflated Poisson and Binomial Regression with Random Effects:A Case Study[J].Biometrics, 2000, 56(4):1030-1039.
[8]王国梁.高速铁路特殊咽喉区接触网布置研究[J].铁道工程学报,2017(10):73-77.
[9]杨桢.接触网分段绝缘器的动态特性分析及维护[J].电气化铁道,2011(2):1-4.
[10]沈修建,吕继涛.刚性悬挂接触网设计若干技术问题探讨[J].电气化铁道,2004(3):60-62.
[11]李雪昆.故障预测与健康管理技术在地铁列车上的应用[J].城市轨道交通研究,2018(2):105-107.