事件影响下的时间序列多尺度集成预测

doi:10.3969/j.issn.1001-506X.2020.02.18

摘要/Abstract

摘要：

针对时间序列的非线性、非平稳和多尺度特征,考虑到事件对序列结构产生的影响,提出事件影响下的时间序列多尺度集成预测策略。首先,基于经验模态分解将原始序列分解成若干分量,从多个尺度展现序列的基本构成;随后,基于迭代累积平方和实现分量序列的变点检验,从多个尺度判别和获取事件对序列产生的结构性影响;然后,基于干预分析构建事件对不同分量序列的影响模型,据此剔除事件影响,获取净化序列;最后,运用基于粒子群优化的支持向量回归,建立单一尺度的序列预测模型,进而实现事件影响下的时间序列多尺度集成预测。实证分析表明:该策略能够精细辨识事件对序列的多尺度影响,有效建立序列总体及分量的预测模型,与传统方法相比,具有更强的事件辨识能力、自适应建模能力和更高的预测精度。

关键词: 时间序列, 事件影响, 经验模态分解, 多尺度分析, 结构性变点, 集成预测

Abstract:

Considering the nonlinear, nonstationary, multiscale characteristics of a complex time series, as well as the structural influence of events, an integrated prediction strategy based on the multiscale series decomposition is proposed. Firstly, an empirical mode decomposition is used to decompose the original series at multiple scales, which is reduced to its original components. Secondly, an iterated cumulative sum of squares method is used to test the structural change points of the original series and its each component, and the structural influence of events on the series at multiple scales can be identified and obtained. Thirdly, based on the identification of structural change points, an intervention analysis method is used to study the time range and intensity change rule of the event influence at different scales, then the influence of events is eliminated and the purified time series is obtained. Finally, a support vector regression optimized by particle swarm optimization is used to predict the single scale series, the prediction model at each scale is established separately and the integrated prediction on the complex time series under the influence of events is realized. The empirical analysis on the cost series of warship maintenance spare parts shows that this method can effectively obtain the influence of events on the series from a multi-scale perspective and build the prediction model of different scale series adaptively. Compared with the existing methods, it has a stronger ability of event analysis and identification, adaptive modeling and a higher prediction precision.

Key words: time series, event influence, empirical mode decomposition, multiscale analysis, structural change point, integrated prediction

中图分类号:

TK42

蒋铁军, 张怀强, 周成杰. 事件影响下的时间序列多尺度集成预测[J]. 系统工程与电子技术, 2020, 42(2): 390-397.

Tiejun JIANG, Huaiqiang ZHANG, Chengjie ZHOU. Multiscale integrated prediction for time series under event influences[J]. Systems Engineering and Electronics, 2020, 42(2): 390-397.

图/表 12

图1

表1

图2

表2

图3

图4

图5

图6

图7

表3

图8

表4

参考文献 32

1	HE Y , XU Q , WAN J , et al. Electrical load forecasting based on self-adaptive chaotic neural network using Chebyshev map[J]. Neural Computing & Applications, 2018, 29 (7): 603- 612.
2	YANG C , QIAO J , WANG L , et al. Dynamical regularized echo state network for time series prediction[J]. Neural Computing & Applications, 2018, 3-4, 1- 14.
3	ZHANG Y , WANG P , NI T , et al. Wind power prediction based on LS-SVM model with error correction[J]. Advances in Electrical and Computer Engineering, 2017, 17 (1): 3- 8.
4	LIN L , WANG F , XIE X . Random forests-based extreme learning machine ensemble for multi-regime time series prediction[J]. Expert Systems with Applications, 2017, 83 (C): 164- 176.
5	JUAN C R , MIGUEL A R C . A wavelet decomposition approach to crude oil price and exchange rate dependence[J]. Economic Modelling, 2013, (32): 42- 57.
6	HUANG L , WANG J . Forecasting energy fluctuation model by wavelet decomposition and stochastic recurrent wavelet neural network[J]. Neurocomputing, 2018, 309 (2): 70- 82.
7	PANNAKKONG W , SRIBOONCHITTA S , HUYNH V N . An ensemble model of ARIMA and ANN with restricted Boltzmann machine based on decomposition of discrete wavelet transform for time series forecasting[J]. Journal of Systems Science and Systems Engineering, 2018, 27 (5): 690- 708. doi: 10.1007/s11518-018-5390-8
8	ROSTAN P , ROSTAN A . The versatility of spectrum analysis for forecasting financial time series[J]. Journal of Forecasting, 2018, 37 (5): 327- 339.
9	HUANG N E , WU M L , QU W D , et al. Applications of Hilbert Huang transform to non-stationary financial time series analysis[J]. Applied Stochastic Models in Business and Industry, 2003, 19 (3): 245- 268. doi: 10.1002/asmb.501
10	WU Z H , HUANG N E . Ensemble empirical mode decomposition:a noise assisted data analysis method[J]. Advances in Adaptive Data Analysis, 2009, 1 (1): 1- 41. doi: 10.1142/S1793536909000047
11	QIU X , REN Y , SUGANTHAN P N , et al. Empirical mode decomposition based ensemble deep learning for load demand time series forecasting[J]. Applied Soft Computing, 2017, 54, 246- 255. doi: 10.1016/j.asoc.2017.01.015
12	崔焕影, 窦祥胜. 基于EMD-GA-BP与EMD-PSO-LSSVM的中国碳市场价格预测[J]. 运筹与管理, 2018, 27 (7): 137- 147.
	CUI H Y , DOU X S . Carbon price forecasts in Chinese carbon trading market based on EMD-GA-BP and EMD-PSO-LSSVM[J]. Operations Research and Management Science, 2018, 27 (7): 137- 147.
13	刘金培, 郭艺, 陈华友, 等. 基于非结构数据流行学习的碳价格多尺度组合预测[J]. 控制与决策, 2019, 34 (2): 58- 65.
	LIU J P , GUO Y , CHEN H Y , et al. Multi-scale combined forecast of carbon price based on manifold learning of unstructured data[J]. Control and Decision, 2019, 34 (2): 58- 65.
14	ANG J S , ZHANG S . Evaluating long-horizon event study methodology[M]. New York: Springer, 2015: 383- 411.
15	CHOW G C . Tests of equality between sets of coefficients in two linear regressions[J]. Econometrica, 1960, 28 (3): 591- 605. doi: 10.2307/1910133
16	BAI J , PERRON P . Estimating and testing linear models with multiple structural changes[J]. Econometrica, 1998, 66, 47- 78. doi: 10.2307/2998540
17	INCLAN C , TIAO G C . Use of cumulative sums of squares for retrospective detection of changes in variance[J]. Journal of the American Statistical Association, 1994, 427 (89): 913- 923.
18	ZHU B , CHEVALLIER J , WEI S M Y . Examining the structural changes of European carbon futures price 2005-2012[J]. Applied Economics Letters, 2015, 22 (5): 335- 342. doi: 10.1080/13504851.2014.943875
19	WEN F , YANG X , GONG X , et al. Multi-scale volatility feature analysis and prediction of gold price[J]. International Journal of Information Technology & Decision Making, 2017, 16 (1): 205- 223.
20	SORESCU A , WARREN N L , ERTEKIN L . Event study methodology in the marketing literature: an overview[J]. Journal of the Academy of Marketing Science, 2017, 45 (2): 186- 207. doi: 10.1007/s11747-017-0516-y
21	FULLER P , YAMANI E , YU G . The impact of the new real estate sector on REITs: an event study[J]. Journal of Economics & Finance, 2017, 43 (4): 1- 19.
22	WOOD L C , WANG J X . The event study method in logistics research[J]. International Journal of Applied Logistics, 2018, 8 (1): 57- 79. doi: 10.4018/IJAL.2018010104
23	DING L , LAM H K S , CHENG T C E , et al. A review of short-term event studies in operations and supply chain management[J]. International Journal of Production Economics, 2018, 200, 329- 342. doi: 10.1016/j.ijpe.2018.04.006
24	MNASRI N E A . Event-study volatility and bootstrapping: an international study[J]. Applied Economics Letters, 2015, 22 (3): 209- 213. doi: 10.1080/13504851.2014.934423
25	BOX G E P , TIAO G C . Intervention analysis with applications to economic and environmental problems[J]. Journal of the American Statistical Association, 1975, 349 (70): 70- 79.
26	LAM C Y , IP W H , LAU C W . A business process activity model and performance measurement using a time series ARIMA intervention analysis[J]. Expert Systems with Applications, 2009, 36 (3): 6986- 6994. doi: 10.1016/j.eswa.2008.08.027
27	RASHED Y , MEERSMAN H , VOORDE E V D , et al. Short-term forecast of container throughout: an ARIMA-intervention model for the port of Antwerp[J]. Maritime Economics & Logistics, 2017, 19 (4): 749- 764.
28	LEUWER D , SVSSMUTH B . Assessing temporary product specific subsidies: a time series intervention analysis[J]. Social Science Electronic Publishing, 2018.
29	MOUSTAKAS A . Assessing the predictive causality of individual based models using Bayesian inference intervention analysis: an application in epidemiology[J]. Stochastic Environmental Research & Risk Assessment, 2018, 32 (10): 2861- 2869.
30	房振明, 王春峰. ICSS方法的中国股市波动结构分析[J]. 哈尔滨工业大学学报, 2009, 41 (10): 295- 298. doi: 10.3321/j.issn:0367-6234.2009.10.067
	FANG Z M , WANG C F . Volatility structure of Chinese stock market based on ICSS method[J]. Journal of Harbin Institute of Technology, 2009, 41 (10): 295- 298. doi: 10.3321/j.issn:0367-6234.2009.10.067
31	MATILLA-GARCIA M , QUERALT R , SANZ P , et al. A generalized BDS statistic[J]. Computational Economics, 2004, 24, 277- 300. doi: 10.1007/s10614-004-4657-y
32	曹小平, 路广安. 装备维修器材保障[M]. 北京: 国防大学出版社, 2005.
	CAO X P , LU G A . Equipment maintenance material support[M]. Beijing: National Defense University Press, 2005.

序号	经费/万元	序号	经费/万元	序号	经费/万元	序号	经费/万元	序号	经费/万元	序号	经费/万元	序号	经费/万元	序号	经费/万元
1	375.62	13	473.33	25	603.76	37	702.20	49	764.03	61	775.99	73	713.51	85	699.89
2	342.25	14	519.03	26	620.72	38	740.39	50	743.64	62	796.85	74	693.12	86	683.48
3	343.92	15	538.31	27	662.25	39	727.70	51	713.33	63	784.24	75	690.06	87	726.12
4	386.47	16	559.17	28	647.32	40	732.14	52	685.79	64	769.13	76	707.12	88	763.48
5	401.11	17	590.13	29	655.95	41	724.54	53	695.90	65	779.70	77	719.35	89	761.44
6	423.45	18	631.19	30	671.89	42	726.40	54	685.15	66	741.41	78	724.64	90	752.45
7	458.87	19	657.43	31	633.70	43	704.89	55	715.09	67	694.14	79	759.40	91	759.77
8	492.42	20	655.20	32	657.98	44	712.40	56	734.18	68	642.23	80	724.64	92	787.86
9	466.37	21	635.92	33	661.32	45	663.36	57	743.92	69	652.05	81	693.77	93	804.73
10	467.67	22	643.99	34	632.40	46	664.66	58	757.36	70	678.75	82	675.97	94	761.07
11	461.65	23	569.18	35	637.68	47	689.87	59	754.49	71	679.77	83	685.05	95	756.43
12	444.68	24	568.25	36	658.45	48	732.42	60	748.00	72	710.64	84	692.47	96	804.17

序列	含义	变点
经费	原始序列	无
IMF1	短期随机波动	无
IMF2	短期波动	无
IMF3	事件影响	22, 69
余项	发展趋势	无

类别	IMF1	IMF2	IMF3	余项
MAE	0.581 4	1.019 9	0.832 6	5.620 5
MSE	0.061 3	0.110 2	0.102 3	0.720 0
MAPE	0.093 0	0.140 9	0.070 5	0.009 0
MSPE	0.015 6	0.056 7	0.014 8	0.001 2

类别	MAE	MSE	MAPE	MSPE
PSO-SVR	8.047 3	3.285 3	0.010 3	0.004 2
EMD-PSO-SVR	4.170 5	1.902 2	0.005 4	0.002 5
EMD-GA-BP	21.590 5	9.877 1	0.027 8	0.012 7
EMD-ARIMA-PLS-NN	15.062 5	6.682 3	0.019 5	0.008 7
本文方法	2.342 6	1.331 4	0.003 0	0.001 7

[1]	张心宇, 刘源, 宋佳凝. 基于LSTM神经网络的短期轨道预报[J]. 系统工程与电子技术, 2022, 44(3): 939-947.
[2]	孙世岩, 张钢, 梁伟阁, 佘博, 田福庆. 基于时间序列数据扩增和BLSTM的滚动轴承剩余寿命预测方法[J]. 系统工程与电子技术, 2022, 44(3): 1060-1068.
[3]	金涛, 王晓峰, 田润澜, 张歆东. 基于改进1DCNN+TCN的雷达辐射源快速识别方法[J]. 系统工程与电子技术, 2022, 44(2): 463-469.
[4]	刘歌, 芮国胜, 田文飚. 基于双重正则矩阵分解的缺失数据恢复[J]. 系统工程与电子技术, 2021, 43(5): 1191-1197.
[5]	崔邦彦, 田润澜, 王东风, 崔钢, 石静苑. 基于注意力机制和改进CLDNN的雷达辐射源识别[J]. 系统工程与电子技术, 2021, 43(5): 1224-1231.
[6]	田睿, 董绪荣. 小波分解与Prophet框架融合的电离层VTEC预报模型[J]. 系统工程与电子技术, 2021, 43(3): 610-622.
[7]	孙云柯, 方志耕, 陈顶. 基于动态灰色主成分分析的多时刻威胁评估[J]. 系统工程与电子技术, 2021, 43(3): 740-746.
[8]	时艳玲, 刘子鹏, 张学良, 顾为亮. 基于EMD能量占比的海面漂浮小目标特征检测[J]. 系统工程与电子技术, 2021, 43(2): 300-310.
[9]	彭正红, 杨德贵, 王行, 王浩, 朱政亮. 基于趋势估计的微多普勒分离与特征提取算法[J]. 系统工程与电子技术, 2021, 43(12): 3452-3461.
[10]	高诗飏, 董会旭, 田润澜, 张歆东. 基于SRNN+Attention+CNN的雷达辐射源信号识别方法[J]. 系统工程与电子技术, 2021, 43(12): 3502-3509.
[11]	孙艺聪, 田润澜, 王晓峰, 董会旭, 戴普. 基于改进CLDNN的辐射源信号识别[J]. 系统工程与电子技术, 2021, 43(1): 42-47.
[12]	张丹娜, 钱锋, 冯辉, 闻年成. Walsh码软扩频信号降噪算法[J]. 系统工程与电子技术, 2020, 42(4): 773-780.
[13]	宋星, 贾红丽, 王谦, 赵汝东. 基于时间序列挖掘的合成旅装备维修保障能力预测[J]. 系统工程与电子技术, 2020, 42(4): 878-886.
[14]	安影, 范训礼, 陈莉, 刘佩. 结合FABEMD和改进的显著性检测的图像融合[J]. 系统工程与电子技术, 2020, 42(2): 292-300.
[15]	王泊涵, 吴超, 柯文俊, 郑恺之, 付修锋, 江山. 基于概率图的作战任务智能规划方法[J]. 系统工程与电子技术, 2020, 42(12): 2795-2801.