间歇性时间序列的可预测性评估及联合预测方法

doi:10.11772/j.issn.1001-9081.2021071196

《计算机应用》唯一官方网站 ›› 2022, Vol. 42 ›› Issue (9): 2722-2731.DOI: 10.11772/j.issn.1001-9081.2021071196

• 数据科学与技术 • 上一篇

间歇性时间序列的可预测性评估及联合预测方法

郎祎平¹, 毛文涛¹^,²(), 罗铁军³, 范黎林¹, 任颖莹⁴, 刘侠¹

^1.河南师范大学计算机与信息工程学院，河南新乡 453007
^2.智慧商务与物联网技术”河南省工程实验室（河南师范大学），河南新乡 453007
^3.株洲中车时代电气股份有限公司，湖南株洲 412001
^4.盾构及掘进技术国家重点实验室，郑州 450001

收稿日期:2021-07-08 修回日期:2021-09-18 接受日期:2021-09-22 发布日期:2021-09-30 出版日期:2022-09-10
通讯作者: 毛文涛
作者简介:郎祎平（1997—），女，河南开封人，硕士研究生，主要研究方向：时间序列预测、工业大数据；
罗铁军（1970—），男，湖南湘潭人，高级工程师，主要研究方向：工业工程；
范黎林（1970—），男，河南周口人，副教授，博士，主要研究方向：工业大数据；
任颖莹（1985—），女，河南南阳人，高级工程师，硕士，主要研究方向：工业大数据；
刘侠（1996—），男，江西赣州人，硕士研究生，主要研究方向：时间序列预测、库存优化。
基金资助:
国家重点研发计划项目(2018YFB1701400);国家自然科学基金资助项目(U1704158);河南省科技攻关项目(212102210103)

Predictability evaluation and joint forecasting method for intermittent time series

Yiping LANG¹, Wentao MAO¹^,²(), Tiejun LUO³, Lilin FAN¹, Yingying REN⁴, Xia LIU¹

^1.College of Computer and Information Engineering，Henan Normal University，Xinxiang Henan 453007，China
^2.Engineering Lab of Intelligence Business and Internet of Things of Henan Province （Henan Normal University），Xinxiang Henan 453007，China
^3.Zhuzhou CRRC Times Electronic Company Limited，Zhuzhou Hunan 412001，China
^4.State Key Laboratory of Shield and Tunneling Technology，Zhengzhou Henan 450001，China

Received:2021-07-08 Revised:2021-09-18 Accepted:2021-09-22 Online:2021-09-30 Published:2022-09-10
Contact: Wentao MAO
About author:LANG Yiping， born in 1997， M. S. candidate. Her research interests include time series forecasting， industrial big data.
LUO Tiejun， born in 1970， senior engineer. His research interests include industrial engineering.
FAN Lilin， born in 1970， Ph. D.， associate professor. His research interests include industrial big data.
REN Yingying， born in 1985， M. S.， senior engineer. Her research interests include industrial big data.
LIU Xia， born in 1996， M. S. candidate. His research interests include time series forecasting， inventory optimization.
Supported by:
National Key Research and Development Program of China(2018YFB1701400);National Natural Science Foundation of China(U1704158);Henan Province Science and Technology Research Project(212102210103)

摘要/Abstract

摘要：

在高端制造企业的运维业务中，配件需求随机发生，且伴随有大量的零需求阶段，同时，对应的配件需求数据量小，且呈现出间歇性和块状分布的特点，导致现有时间序列预测方法难以有效预测配件需求走势。为解决该问题，提出了一种间歇性时间序列的可预测性评估及联合预测方法。首先，提出了一种新的间歇相似度指标，通过统计两条序列中“0”元素出现的频次和位置，并结合最大信息系数和平均需求间隔等度量指标，有效评估了序列的趋势信息和波动规律，并实现了对间歇性序列可预测性的量化；其次，基于该指标，构建了一个间歇相似度层次聚类方法来自适应地筛选相似性高、可预测性强的序列，剔除极度稀疏、无法预测的序列；此外，探索利用序列间的结构化信息，并构建多输出支持向量回归（M-SVR）模型，从而实现小样本下的间歇性序列联合预测；最后，分别在两个公开数据集（UCI礼品零售数据集和华为电脑配件数据集）和某大型制造企业实际配件售后数据集上进行实验。实验结果表明，相比多个典型的时间序列预测方法，所提方法可有效挖掘各类间歇性序列的可预测性，提高小样本间歇性序列的预测精度，从而为制造企业配件需求预测提供了一种新的解决方案。

关键词: 需求预测, 间歇性时间序列, 可预测性评估, 时间序列预测, 时间序列聚类

Abstract:

In the operation and maintenance of high-end manufacturing enterprises， the spare parts demand occurs randomly， accompanied by a large number of zero demand periods. At the same time， the corresponding sparse parts demand data is of small scale and has intermittent and distribution with lump formation characteristics. Consequently， most of current time series forecasting methods are hard to effectively predict the demand trends. To solve this problem， a predictability evaluation and joint forecasting method for intermittent time series was proposed. Firstly， a new intermittent-similarity metric was proposed. In this metric， the frequency and positions of the "0" element occurring in the two sequences were counted， while the metrics such as maximal information coefficient and average demand interval were combined to evaluate the tendency information and fluctuation pattern of the sequences effectively and realize the quantification of the predictability of the intermittent time series. Then， based on this metric， an intermittent-similarity hierarchical clustering method was constructed to adaptively select the sequences with high similarity and strong predictability as well as eliminate extremely sparse and unpredictable sequences. Moreover， the structured information between the sequences was explored and utilized， a Multi-output Support Vector Regression （M-SVR） model was constructed， thereby achieving the joint prediction of intermittent time series with small-scale data. Finally， the experiments were conducted on two public datasets （UCI （University of California Irvine） gift retail dataset and Huawei computer accessory dataset） and a real-world spare parts after-sales dataset of a large manufacturing enterprise， respectively. The results show that compared with several representative time series forecasting methods， the proposed method can effectively exploit the predictability of various kinds of intermittent sequences and improve the prediction accuracy of intermittent time series with small-scale data. Therefore， the proposed method provides a new solution for the spare parts demand forecasting of manufacturing enterprises.

Key words: demand forecasting, intermittent time series, predictability evaluation, time series forecasting, time series clustering

中图分类号:

TP181

郎祎平, 毛文涛, 罗铁军, 范黎林, 任颖莹, 刘侠. 间歇性时间序列的可预测性评估及联合预测方法[J]. 计算机应用, 2022, 42(9): 2722-2731.

Yiping LANG, Wentao MAO, Tiejun LUO, Lilin FAN, Yingying REN, Xia LIU. Predictability evaluation and joint forecasting method for intermittent time series[J]. Journal of Computer Applications, 2022, 42(9): 2722-2731.

图/表 18

图1 本文方法流程

Fig. 1 Flowchart of the proposed method

图2 间歇性序列 MIC 计算示意图

Fig. 2 Schematic diagram of MIC calculation of intermittent sequences

表1 实验数据集信息

Tab. 1 Experimental dataset information

数据集

数据

长度

数据

维度

属性

数目

配件编号、时间（月）、每月故障数量、每月修复入库数量、每月故障出库数量等

图3 本文实验数据走势分布情况

Fig. 3 Trend distribution of experimental data in this paper

图4 本文实验数据的ADI和CV2分布情况

Fig. 4 Distribution of ADI and CV2 in experimental data

表2 对比方法

Tab. 2 Comparison methods

算法类型	对比方法
单维时间序列预测	ARIMA^［3］
单维时间序列预测	SVR^［4］
多元时间序列联合预测	ELM^［21］
	LSTM^［8］
	BP神经网络^［7］
	BHT-ARIMA^［16］

图5 华为电脑配件数据集部分聚类结果示例

Fig. 5 Some clustering result examples of Huawei computer accessory dataset

图6 UCI礼品零售数据集部分聚类结果示例

Fig. 6 Some clustering result examples of UCI gift retail dataset

图7 华为电脑配件数据集上不同方法预测结果对比

Fig. 7 Prediction results comparison by different methods on Huawei computer accessory dataset

图8 UCI礼品零售数据集上不同方法预测结果对比

Fig. 8 Prediction results comparison by different methods on UCI gift retail dataset

图9 公开数据集上不同预测方法的性能对比

Fig. 9 Performance comparison of different prediction methods on public datasets

图10 华为电脑配件数据集上结果异常的序列分析

Fig. 10 Analysis of sequences with abnormal results on Huawei computer accessory dataset

图11 公开数据集上不同参数取值所对应的MSE

Fig. 11 MSE corresponding to different parameter values on public datasets

图12 实际配件售后数据集上部分聚类结果示例

Fig. 12 Some clustering result examples on real-world spare parts after-sales dataset

图13 本文方法在21号等配件序列的预测结果

Fig. 13 Prediction results of the proposed method on sequences such as No. 21

图14 本文方法在127号等配件序列上的预测结果

Fig. 14 Prediction results of the proposed method on sequences such as No.127

图15 实际配件售后数据集上不同方法预测结果对比

Fig. 15 Comparison of prediction results by different methods on real-world spare parts after-sales dataset

图16 实际配件售后数据集上不同预测方法性能对比

Fig. 16 Performance comparison of different prediction methods on real-world spare parts after-sales dataset

参考文献 21

1	BAO Y K， WANG W， ZOU H. SVR-based method forecasting intermittent demand for service parts inventories［C］// Proceedings of the 2005 International Workshop on Rough Sets， Fuzzy Sets， Data Mining， and Granular Computing， LNCS 3642. Berlin： Springer， 2005： 604-613.
2	CROSTON J D. Forecasting and stock control for intermittent demands［J］. Journal of the Operational Research Society， 1972， 23（3）： 289-303. 10.1057/jors.1972.50
3	CHEN Y， ZHAO H， YU L. Demand forecasting in automotive aftermarket based on ARMA model［C］// Proceedings of the 2010 International Conference on Management and Service Science. Piscataway： IEEE， 2010： 1-4. 10.1109/icmss.2010.5577867
4	KARMY J P， MALDONADO S. Hierarchical time series forecasting via support vector regression in the European travel retail industry［J］. Expert Systems with Applications， 2019， 137： 59-73. 10.1016/j.eswa.2019.06.060
5	WANG Y， GUO Y K. Forecasting method of stock market volatility in time series data based on mixed model of ARIMA and XGBoost［J］. China Communications， 2020， 17（3）： 205-221. 10.23919/jcc.2020.03.017
6	VAN STEENBERGEN R M， MES M R K. Forecasting demand profiles of new products［J］. Decision Support Systems， 2020， 139： No.113401. 10.1016/j.dss.2020.113401
7	XING R R， SHI X L. A BP-SVM combined model for intermittent spare parts demand prediction［C］// Proceedings of the 2019 IEEE International Conference on Systems， Man， and Cybernetics. Piscataway： IEEE， 2019： 1085-1090. 10.1109/smc.2019.8914609
8	ABBASIMEHR H， SHABANI M， YOUSEFI M. An optimized model using LSTM network for demand forecasting［J］. Computers and Industrial Engineering， 2020， 143： No.106435. 10.1016/j.cie.2020.106435
9	CAO J S， WANG J H. Stock price forecasting model based on modified convolution neural network and financial time series analysis［J］. International Journal of Communication Systems， 2019， 32（12）： No.e3987. 10.1002/dac.3987
10	MOR R S， NAGAR J， BHARDWAJ A. A comparative study of forecasting methods for sporadic demand in an auto service station［J］. International Journal of Business Forecasting and Marketing Intelligence， 2019， 5（1）： 56-70. 10.1504/ijbfmi.2019.099009
11	BOUKHTOUTA A， JENTSCH P. Support vector machine for demand forecasting of Canadian Armed Forces spare parts［C］// Proceedings of the 6th International Symposium on Computational and Business Intelligence. Piscataway： IEEE， 2018： 59-64. 10.1109/iscbi.2018.00021
12	FU W H， CHEN C F， LIN Z H. A hybrid forecasting framework with neural network and time-series method for intermittent demand in semiconductor supply chain［C］// Proceedings of the 2018 IFIP International Conference on Advances in Production Management Systems， IFIPAICT 536. Cham： Springer， 2018： 65-72.
13	TEUNTER R H， DUNCAN L. Forecasting intermittent demand： a comparative study［J］. Journal of the Operational Research Society， 2009， 60（3）： 321-329. 10.1057/palgrave.jors.2602569
14	XU S J， CHAN H K， CH’NG E， et al. A comparison of forecasting methods for medical device demand using trend-based clustering scheme［J］. Journal of Data， Information and Management， 2020， 2（2）： 85-94. 10.1007/s42488-020-00026-y
15	BAO Y K， WANG W， ZHANG J L. Forecasting intermittent demand by SVMs regression［C］// Proceedings of the 2004 IEEE International Conference on Systems， Man and Cybernetics. Piscataway： IEEE， 2004： 461-466. 10.1109/icsmc.2004.1400628
16	SHI Q Q， YIN J M， CAI J J， et al. Block Hankel tensor ARIMA for multiple short time series forecasting［C］// Proceedings of the 34th AAAI Conference on Artificial Intelligence. Palo Alto， CA： AAAI Press， 2020： 5758-5766. 10.1609/aaai.v34i04.6032
17	陈美云. 时间序列聚类分析中几种算法的研究及应用［D］. 徐州：中国矿业大学， 2019：7-8.
	CHEN M Y. Comparing clustering algorithm using time-series data［D］. Xuzhou： China University of Mining and Technology， 2019： 7-8.
18	毛文涛，赵胜杰，张俊娜. 基于主曲线的多输入多输出支持向量机算法［J］. 计算机应用， 2013， 33（5）： 1281-1284， 1293. 10.3724/sp.j.1087.2013.01281
	MAO W T， ZHAO S J， ZHANG J N. Multi-input-multi-output support vector machine based on principal curve［J］. Journal of Computer Applications， 2013， 33（5）： 1281-1284， 1293. 10.3724/sp.j.1087.2013.01281
19	RESHEF D N， RESHEF Y A， FINUCANE H K， et al. Detecting novel associations in large data sets［J］. Science， 2011， 334（6062）： 1518-1524. 10.1126/science.1205438
20	TÜRKMEN A C， WANG Y Y， JANUSCHOWSKI T. Intermittent demand forecasting with deep renewal processes［EB/OL］. （2019-11-23）［2021-07-08］.. 10.1371/journal.pone.0259764
21	MAO W T， ZHAO S J， MU X X， et al. Multi-dimensional extreme learning machine［J］. Neurocomputing， 2015， 149（Pt A）：160-170. 10.1016/j.neucom.2014.02.073

[1]	吴明晖, 张广洁, 金苍宏. 基于多模态信息融合的时间序列预测模型[J]. 《计算机应用》唯一官方网站, 2022, 42(8): 2326-2332.
[2]	王艺霏, 于雷, 滕飞, 宋佳玉, 袁玥. 基于长-短时序特征融合的资源负载预测模型[J]. 《计算机应用》唯一官方网站, 2022, 42(5): 1508-1515.
[3]	胡鹤轩, 隋华超, 胡强, 张晔, 胡震云, 马能武. 基于图注意力网络与双阶注意力机制的径流预报模型[J]. 《计算机应用》唯一官方网站, 2022, 42(5): 1607-1615.
[4]	李洋莹, 陈智军, 张子豪, 游兰. 基于改进Elman神经网络的制糖企业原糖需求预测模型[J]. 计算机应用, 2021, 41(7): 2113-2120.
[5]	肖勇, 郑楷洪, 郑镇境, 钱斌, 李森, 马千里. 基于多尺度跳跃深度长短期记忆网络的短期多变量负荷预测[J]. 计算机应用, 2021, 41(1): 231-236.
[6]	徐雅斌, 彭宏恩. 基于需求预测的PaaS平台资源分配方法[J]. 计算机应用, 2019, 39(6): 1583-1588.
[7]	王金策, 邓越萍, 史明, 周云飞. 多时间尺度时间序列趋势预测[J]. 计算机应用, 2019, 39(4): 1046-1052.
[8]	周翔宇, 程勇, 王军. 基于改进深度信念网络的农业温室温度预测方法[J]. 计算机应用, 2019, 39(4): 1053-1058.
[9]	王坤, 员晓阳, 王力. 基于改进模糊支持向量回归模型的机场能源需求预测[J]. 计算机应用, 2016, 36(5): 1458-1463.
[10]	李瑞国, 张宏立, 王雅. 基于量子粒子群优化算法的新型正交基神经网络分数阶混沌时间序列单步预测[J]. 计算机应用, 2015, 35(8): 2227-2232.
[11]	郑伟谷大武陆海宁. 改进型AR(1)模型在域名系统中的应用[J]. 计算机应用, 2010, 30(3): 736-739.
[12]	陈云亮杨捷杨娟李德华康立山. 基于均匀设计的GEP算法研究与应用[J]. 计算机应用, 2007, 27(4): 948-951.
[13]	丁宁周新志 . 基于提升方案的冗余Haar小波变换与时间序列预测[J]. 计算机应用, 2007, 27(1): 58-60.
[14]	韩雪梅; 徐从富; 沈慧峰. 基于序列分解的复杂系统的时序预测方法[J]. 计算机应用, 2006, 26(4): 888-890.

间歇性时间序列的可预测性评估及联合预测方法

Predictability evaluation and joint forecasting method for intermittent time series

RichHTML

PDF

可视化

摘要/Abstract

引用本文

使用本文

图/表 18

参考文献 21

相关文章 14

编辑推荐

Metrics