摘要
为了解决传统索赔监控与预警模型存在的未考虑索赔率大小对预警能力的影响、模型对保证数据精度要求高、难以处理索赔率较低的失效模式等问题,基于非参数估计和泊松分布对单位时间内产品保证索赔记录进行了拟合,根据各监控阶段的可用信息量大小构造了分阶段风险调整控制图对保证索赔进行监控和预警,并采用了在指定监控时间内对所有销售产品进行统一监控的监控方式。模型的有效性利用实际算例从误报警率、报警能力和响应速度三个指标进行了分析。
In order to solve the problem that traditional product claim monitoring and early warning model had not consider the influence of claim rate to the warning ability, as well as the model need accurate warranty data and had difficult to deal with failure modes with low claim rate, this paper simulates the product warranty claims records in unit time based on the parameter estimation and Poisson distribution.According to the available information of each monitoring stage, risk-adjusted control chart was constructed to monitor the warranty claims and unified monitoring method was adopted in a specified monitoring time to all sales products. The model is validated through the analysis actual example from three aspects:the false alarm rate, alarm ability and response speed.
出处
《数理统计与管理》
CSSCI
北大核心
2017年第3期496-505,共10页
Journal of Applied Statistics and Management
基金
国家自然科学基金资助项目(71171142)
高等学校博士学科点专项科研基金资助项目(20110032110034)
关键词
保证索赔
可靠性
风险调整
控制图
安全预警
warranty claim, reliability, risk-adjusted, control chart, early warning