基于二维保修数据的汽车索赔及保修费用预测方法

doi:10.3969/j.issn.1674-8484.2023.03.006

摘要/Abstract

摘要：

为了提高汽车索赔率预测精度并准确预测车企在未来时段的保修费用，考虑二维保修数据的删失特性和汽车的过保情形，提出了基于在保率修正的汽车索赔率计算方法。鉴于汽车保修时零件维修和更换两种模式下保修费用的差异，提出了基于索赔率和维修更换率的保修费用预测方法。结合某品牌汽车的实际二维保修数据，预测了车企未来一个季度的保修费用。结果表明：预测的保修费用与实际保修费用的相对误差在6.5%以内；相较于不考虑汽车在保率的方法，所提保修费用预测方法的预测精度提高了4.5%以上，可以为车企保修策略制定、产品定价及保修准备金规划等提供支撑。

关键词: 汽车保修, 二维保修数据, 索赔率, 保修费用

Abstract:

A claim ratio calculation method based on the correction of in-warranty ratio was proposed to improve the prediction accuracy of automobile claim ratio and accurately predict the warranty cost of automobile enterprises in the future period, considering the censoring characteristics of two-dimensional warranty data and the over-warranty situation of automobile products. A prediction method of the warranty cost based on the claim ratio and repair-replacement rate was proposed due to the difference of the warranty cost under the two modes of parts repair and replacement. The warranty cost of the automobile enterprise in the next quarter was predicted, combining with the actual two-dimensional warranty data of a certain brand of automobile. The results show that the relative error between the predicted warranty cost and the actual warranty cost is within 6.5%. Compared with the method that does not consider the in-warranty ratio of automobile, the prediction accuracy of the proposed warranty cost prediction method is improved by more than 4.5%, which can provide support for warranty strategies formulation, product pricing and warranty reserve planning of automobile enterprises.

Key words: automobile warranty, two-dimensional warrantydata, claim ratio, warranty costs

中图分类号:

TH17
TB114.3

庞欢, 石东阳, 宫政, 刘敬一. 基于二维保修数据的汽车索赔及保修费用预测方法[J]. 汽车安全与节能学报, 2023, 14(3): 310-318.

PANG Huan, SHI Dongyang, GONG Zheng, LIU Jingyi. Prediction of automobile claims and warranty costs method based on two-dimensional warranty data[J]. Journal of Automotive Safety and Energy, 2023, 14(3): 310-318.

图/表 12

参考文献 29

[1]	LIN Kunsong, CHEN Yunxia. Analysis of two-dimensional warranty data considering global and local dependence of heterogeneous marginal[J]. Reli Engi Syst Safe, 2021, 207: 107-116.
[2]	WU Shaomin. Warranty Data Analysis: A Review[J]. Qual Reli Engi Int’l, 2012, 28(8): 795-805.
[3]	戴安舒, 魏冠州, 何曙光, 等. 二维质保管理研究综述与展望[J]. 工业工程, 2019, 22(6): 96-102+117.
	DAI Anshu, WEI Guanzhou, HE Shuguang, et al. Two dimensional warranty management review and research prospect[J]. Ind Engi J, 2019, 22(6): 96-102+117.. (in Chinese)
[4]	揭丽琳, 刘卫东. 基于使用可靠性的产品区域保修差别定价策略系统动力学模型[J]. 系统工程理论与实践, 2019, 39(1): 236-250. doi: 10.12011/1000-6788-2018-0265-15
	JIE Lilin, LIU Weidong. System dynamics model for regional warranty differential pricing strategy based on operational reliability[J]. Syst Engi-Theo Pract, 2019, 39(1): 236-250.. (in Chinese)
[5]	戴安舒, 何桢, 刘子先, 等. 二维保修产品的柔性预防性维护策略研究[J]. 工业工程与管理, 2017, 22(5): 9-16.
	DAI Anshu, HE Zhen, LIU Zixian, et al. Study on the flexible preventive maintenance strategy for products sold with two-dimensional warranty[J]. Ind Engi Manag, 2017, 22(5): 9-16.. (in Chinese)
[6]	SHANG Lijun, QIU Qingan, WANG Xin. Random periodic replacement models after the expiry of 2D-warranty[J]. Comput Ind Engi, 2022, 164: 107885.
[7]	SHANG Lijun, SI Shubin, CAI Zhiqiang. Optimal maintenance-replacement policy of products with competing failures after expiry of the warranty[J]. Comput Ind Engi, 2016, 98: 68-77.
[8]	XIE Wei, YE Zhisheng. Aggregate discounted warranty cost forecast for a new product considering stochastic sales[J]. IEEE Trans Reli, 2016, 65(1): 486-497.
[9]	CAI Keyuan, HE Shuguang, HE Zhen. Information sharing under different warranty policies with cost sharing in supply chains[J]. Int’l Trans Oper Res, 2020, 27(3): 1550-1572.
[10]	Majeske K D. A non-homogeneous Poisson process predictive model for automobile warranty claims[J]. Reli Engi Syst Safe, 2007, 92(2): 243-251.
[11]	DAI Anshu, ZHANG Zhaomin, HOU Pengwen, et al. Warranty claims forecasting for new products sold with a two-dimensional warranty[J]. J Syst Sci Syst Engi, 2019, 28(6): 715-730.
[12]	WU Shaomin, Akbarov A. Forecasting warranty claims for recently launched products[J]. Reli Engi Syst Safe, 2012, 106: 160-164.
[13]	WU Shaomin. A review on coarse warranty data and analysis[J]. Reli Engi Syst Safe, 2013, 144: 1-11.
[14]	杜文超, 何曙光, 边德军, 等. 二维保修下基于使用强度的备件库存策略研究[J]. 工业工程与管理, 2017, 22(2): 119-124.
	DU Wenchao, HE Shuguang, BIAN Dejun, et al. Inventory control policy for parts concerning the usage rate’s effect to product reliability of two-dimensional warranty[J]. Ind Engi Manag, 2017, 22(2): 119-124.. (in Chinese)
[15]	Hrycej T, Grabert M. Warranty cost forecast based on car failure data[C]// Proceed Int’l Joint Conf Neural Networks. Orlando, Florida, USA: IEEE Press, 2007, 6: 108-113.
[16]	Rai B. Warranty spend forecasting for subsystem failures influenced by calendar month seasonality[J]. IEEE Trans Reli, 2009, 58(4): 649-657.
[17]	WANG Xin, XIE Wei, YE Zhisheng, et al. Aggregate discounted warranty cost forecasting considering the failed-but-not-reported events[J]. Reli Engi Syst Safe, 2017, 168: 355-364.
[18]	马小敏, 何曙光, 李婷. 考虑用户更换行为的质保成本预测[J]. 工业工程, 2022, 25(2): 163-167.
	MA Xiaomin, HE Shuguang, LI Ting. Warranty cost prediction considering users’ replacement behavior[J]. Ind Engi J, 2022, 25(2):163-167.. (in Chinese)
[19]	SHANG Lijun, SI Shubin, SUN Shudong, et al. Optimal warranty design and post-warranty maintenance for products subject to stochastic degradation[J]. IISE Trans, 2018, 50(10): 913-927. doi: 10.1080/24725854.2018.1448490 URL
[20]	SHANG Lijun, LIU Baoliang, QIU Qingan, et al. Designing warranty and maintenance policies for products subject to random working cycles[J]. Reli Engi Syst Safe, 2023, 234: 109187.
[21]	庞欢, 石东阳, 王道成, 等. 基于质保数据的汽车可靠性评估及车企运营决策系统[J]. 汽车安全与节能学报, 2022, 13(2): 250-258.
	PANG Huan, SHI Dongyang, WANG Daocheng, et al. Automobile reliability evaluation and automobile-enterprise operational decision system based on warranty data[J]. J Autom Safe Energ, 2022, 13(2): 250-258.. (in Chinese)
[22]	WANG Xiaolin, XIE Wei. Two-dimensional warranty: A literature review[J]. Proceed Institut Mech Engi Part O: J Risk Reli, 2018, 232(3): 284-307.
[23]	董方岐, 刘子先, 郝剑虹, 等. 考虑预防性维修的二维保修期设计与优化研究[J]. 工业工程与管理, 2019, 24(4): 64-71+80.
	DONG Fangqi, LIU Zixian, HAO Jianhong, et al. The designand optimization of two-dimensional warranty period with considering preventive maintenance strategy[J]. Ind Engi Manag, 2019, 24(4): 64-71+80. (in Chinese)
[24]	YANG Duo, HE Zhen, HE Shuguang. Warranty claims forecasting based on a general imperfect repair model considering usage rate[J]. Reli Engi Syst Safe, 2016, 145: 147-154.
[25]	Pej G, Bijelic F, Zovak G. Model for Calculating average vehicle mileage for different vehicle classes based on real data: A case study of Croatia[J]. Promet-Zagreb, 2019, 31(2): 213-222.
[26]	林秀丽, 汤大钢, 丁焰, 等. 中国机动车行驶里程分布规律[J]. 环境科学研究, 2009, 22(3): 337-380.
	LIN Xiuli, TANG Dagang, DING Yan, et al. Study on the distribution of vehicle mileage traveled in China[J]. Res Environ Sci, 2009, 22(3): 377-380.. (in Chinese)
[27]	朱建平, 李治国, 陈彩云. 数据挖掘中一种新的预测模型: 基函数拟合预测及其在股市中的应用[J]. 系统工程理论与实践, 2003(9): 35-40+60. doi: 10.12011/1000-6788(2003)9-35
	ZHU Jianping, LI Zhiguo, CHEN Caiyun. A new predictive model on data mining: Predicting arithmetic of bernstein basic function fitting and its application for stock market[J]. Syst Engi-Theo Pract, 2003(9): 35-40+60. (in Chinese).
[28]	Fleurke S. Forecasting Automobile Sales Using an Ensemble of Methods[J]. WSEAS Trans Syst, 2017, 16: 337-345.
[29]	Claudia C, Rainer Kastenmeier R, Eike C, et al. A mixed copula model for insurance claims and claim sizes[J]. Scand Actu J, 2012, 4: 278-305.

行驶时间 t / 月	总车辆数 N / 辆	在保概率 P	修正后车辆数 N×P
1	N₁	P₁	N₁×P₁
2	N₂	P₂	N₂×P₂
…	…	…	…
n	N_n	P_n	N_n×P_n

行驶时间 t / 月	总车辆数 N / 辆	在保概率 P	修正后车辆数 N×P
1	N₁	P₁	N₁×P₁
2	N₂	P₂	N₂×P₂
…	…	…	…
n	N_n	P_n	N_n×P_n

月区间	索赔数	月区间	索赔数
0~1	6	13~14	54
1~2	12	14~15	50
2~3	26	15~16	58
3~4	24	16~17	43
4~5	36	17~18	40
5~6	40	18~19	50
6~7	63	19~20	35
7~8	48	20~21	2
8~9	70	21~22	25
9~10	54	22~23	15
10~11	78	23~24	8
11~12	62	24~25	1
12~13	77	-	-

月区间	索赔数	月区间	索赔数
0~1	6	13~14	54
1~2	12	14~15	50
2~3	26	15~16	58
3~4	24	16~17	43
4~5	36	17~18	40
5~6	40	18~19	50
6~7	63	19~20	35
7~8	48	20~21	2
8~9	70	21~22	25
9~10	54	22~23	15
10~11	78	23~24	8
11~12	62	24~25	1
12~13	77	-	-

索赔件名称	预测时段	实际值/元	不考虑在保率的方法			考虑在保率的方法
索赔件名称	预测时段	实际值/元	预测值/元	误差/ %	总误差/ %	预测值/元	误差/ %	总误差/ %
制动盘	2020/07	17 457.26	18 774.85	7.55	5.23	17 957.81	2.87	0.08
	2022/08	18 101.65	19 968.34	10.31		18 984.17	4.88
	2022/09	22 459.45	22 311.21	-0.66		21 121.61	-5.96
悬架后上横臂	2022/07	18 101.65	19 322.90	6.75	11.02	18 677.98	3.18
	2022/08	16 909.20	18 469.59	9.23		17 678.79	4.55	6.21
	2022/09	14 598.45	17 281.74	18.38		16 334.99	11.90