Modeling Time Series Using Multi-Modality Fuzzy Cognitive Maps

doi:10.16182/j.issn1004731x.joss.20-0834

Abstract

Abstract:

A multi-modality modeling method for time series data based on fuzzy cognitive maps is proposed to address the problem that a single model is difficult to accurately reflect the multi-modal characteristics of time series.The bootstrap method is used to select multiple sub-sequences from the original time serieswhich contain the diverse modality in the original time series. The fuzzy cognitive map sub-models are constructed on each sub-sequencesrespectively. The formed sub-models are further merged by means of granular computing method and the merging performance with different weighting strategies is analyzed. The developed multi-modal model not only has prediction abilities at the numeric and interval level, but also has better interpretability. Experimental results on public datasets exhibit the usefulness and satisfactory efficiency of the proposed approach.

Key words: fuzzy cognitive maps, granular computing, time series, model weight

CLC Number:

TP274

Guoliang Feng, Wei Lu, Jianhua Yang. Modeling Time Series Using Multi-Modality Fuzzy Cognitive Maps[J]. Journal of System Simulation, 2022, 34(3): 543-554.

Figures/Tables 14

Fig. 1

Fig. 2

Fig. 3

Fig. 4

Table 1

Table 2

Fig. 5

Fig. 6

Fig. 7

Fig. 8

Table 3

Fig. 9

Table 4

Table 5

References 32

1	Weigend Andreas S, Gerschenfeld Neil A. Time Series Prediction: Forecasting the Future and Understanding the Past[J]. International Journal of Forecasting (S0169-2070), 1994, 10(1): 463-466.
2	Lee Y S, Tong L I. Forecasting Time Series Using a Methodology Based on Autoregressive Integrated Moving Average and Genetic Programming[J]. Knowledge-Based Systems (S0950-7051), 2011, 24(1): 66-72.
3	Zhang G P. Time Series Forecasting Using a Hybrid ARIMA and Neural Network Model[J]. Neurocomputing (S0925-2312), 2003, 50: 159-175.
4	Khashei M, Bijari M, Ardali G A R. Hybridization of Autoregressive Integrated Moving Average (ARIMA) with Probabilistic Neural Networks (PNNs)[J]. Computers & Industrial Engineering (S0360-8352), 2012, 63(1): 37-45.
5	Cao L J, Tay F E H. Support vector Machine with Adaptive Parameters in Financial Time Series Forecasting[J]. IEEE Transactions on Neural Networks (S1045-9227), 2003, 14(6): 1506-1518.
6	Chaâbane N. A Novel Auto-Regressive Fractionally Integrated Moving Average-Least-Squares Support Vector Machine Model for Electricity Spot Prices Prediction[J]. Journal of Applied Statistics (S0266-4763), 2014, 41(3): 635-651.
7	Guo H, Liu X, Sun Z. Multivariate Time Series Prediction Using a Hybridization of VARMA Models and Bayesian Networks[J]. Journal of Applied Statistics (S0266-4763), 2016, 43(16): 2897-2909.
8	Song Q, Chissom B S. Fuzzy Time Series and its Models[J]. Fuzzy Sets and Systems (S0165-0114), 1993, 54(3): 269-277.
9	Singh P. A Brief Review of Modeling Approaches Based on Fuzzy Time Series[J]. International Journal of Machine Learning and Cybernetics (S1868-8071), 2017, 8(2): 397-420.
10	Bose M, Mali K. Designing Fuzzy Time Series Forecasting Models: a Survey[J]. International Journal of Approximate Reasoning (S0888-613X), 2019, 111: 78-99.
11	Lu W, Chen X, Pedrycz W, et al. Using Interval Information Granules to Improve Forecasting in Fuzzy Time Series[J]. International Journal of Approximate Reasoning (S0888-613X), 2015, 57: 1-18.
12	Froelich W, Pedrycz W. Fuzzy Cognitive Maps in the Modeling of Granular Time Series[J]. Knowledge-Based Systems (S0950-7051), 2017, 115: 110-122.
13	Bargiela A, Pedrycz W. Toward a Theory of Granular Computing for Human-Centered Information Processing[J]. IEEE Transactions on Fuzzy Systems (S1063-6706), 2008, 16(2): 320-330.
14	Pedrycz W. The Principle of Justifiable Granularity and an Optimization of Information Granularity Allocation as Fundamentals of Granular Computing[J]. Journal of Information Processing Systems (S1976-913X), 2011, 7(3): 397-412.
15	Pedrycz W, Homenda W. Building the Fundamentals of Granular Computing: a Principle of Justifiable Granularity[J]. Applied Soft Computing (S1568-4946), 2013, 13(10): 4209-4218.
16	Papageorgiou E I, Salmeron J L. A Review of Fuzzy Cognitive Maps Research During the Last Decade[J]. IEEE Transactions on Fuzzy Systems (S1063-6706), 2012, 21(1): 66-79.
17	Felix G, Nápoles G, Falcon R, et al. A Review on Methods and Software for Fuzzy Cognitive Maps[J]. Artificial Intelligence Review (S0269-2821), 2019, 52(3): 1707-1737.
18	Stach W, Kurgan L A, Pedrycz W. Numerical and Linguistic Prediction of Time Series with the Use of Fuzzy Cognitive Maps[J]. IEEE Transactions on Fuzzy Systems (S1063-6706), 2008, 16(1): 61-72.
19	Lu W, Yang J, Liu X, et al. The Modeling and Prediction of Time Series Based on Synergy of High-Order Fuzzy Cognitive Map and Fuzzy c-Means Clustering[J]. Knowledge-Based Systems (S0950-7051), 2014, 70: 242-255.
20	Pedrycz W, Jastrzebska A, Homenda W. Design of Fuzzy Cognitive Maps for modeling time series[J]. IEEE Transactions on Fuzzy Systems (S1063-6706), 2016, 24(1): 120-130.
21	Salmeron J L, Froelich W. Dynamic Optimization of Fuzzy Cognitive Maps for Time Series Forecasting[J]. Knowledge-Based Systems (S0950-7051), 2016, 105: 29-37.
22	Froelich W, Pedrycz W. Fuzzy Cognitive Maps in the Modeling of Granular Time Series[J]. Knowledge-Based Systems (S0950-7051), 2017, 115: 110-122.
23	Felix G, N´apoles G, Falcon R, et al. A Review on Methods and Software for Fuzzy Cognitive Maps[J]. Artificial Intelligence Review (S0269-2821) , 2019(52): 1707-1737.
24	Zadeh L A. Toward a Theory of fuzzy Information Granulation and its Centrality in Human Reasoning and Fuzzy Logic[J]. Fuzzy Sets and Systems (S0165-0114), 1997, 90(2): 111-127.
25	Zadeh L A. Is There a Need for Fuzzy Logic?[J]. Information Sciences (S0020-0255), 2008, 178(13): 2751-2779.
26	Bargiela A, Pedrycz W. Toward a Theory of Granular Computing for Human-Centred Information Processing[J]. IEEE Transactions on Fuzzy Systems (S1063-6706), 2008, 16(2):320-330.
27	Pedrycz W. Granular Computing for Data Analytics: A Manifesto of Human-Centric Computing[J]. IEEE/CAA Journal of Automatica Sinica (S2329-9266), 2018, 5(6): 1025-1034.
28	Pedrycz W, Vukovich G. Abstraction and Specialization of Information Granules[J]. IEEE Transactions on Systems, Man and Cybernetics - PartB: Cybernetics (S1083-4419), 2001, 31(1): 106-111.
29	Stach W, Kurgan L A, Pedrycz W. Numerical and Linguistic Prediction of Time Series with the Use of Fuzzy Cognitive Maps[J]. IEEE Transactions on Fuzzy Systems (S1063-6706), 2008, 16(1): 61-72.
30	Boyd S, Vandenberghe L. Convex Optimization[M]. London: Cambridge University Press, 2004.
31	Khosravi A, Nahavandi S, Creighton D. Prediction Intervals for Short-Term Wind Farm Power Generation Forecasts[J]. IEEE Transactions on Sustainable Energy (S1949-3029), 2013, 4(3): 602-610.
32	Quan H, Srinivasan D, Khosravi A. Short-Term Load and Wind Power Forecasting Using Neural Network-Based Prediction Intervals[J]. IEEE Transactions on Neural Networks and Learning Systems (S2162-237X), 2013, 25(2): 303-315.

序号	主要参数	量值
1	λ	5
2	子模型数量p	100
3	节点数n	3
4	子序列长度k	5
5	粒度水平α	1

数据集	幅值低	幅值中	幅值高
MG	0.219 2	0.766 5	1.313 7
AUD/USD汇率	0.598 0	1.043 0	1.488 0
Vatnsdalsa河流量	3.670 0	28.835 0	54.000 0
风速时间序列	0.090 0	9.940 0	19.790 0

数据集	权重方法	PICP	PINAW	CWC	RMSE
MG	动态权重	0.912 5	0.147 2	0.201 1	0.026 5
	模型权重	0.920 8	0.292 5	0.399 8	0.116 9
	平均权重	0.950 0	0.357 1	0.488 3	0.107 4
汇率	动态权重	0.873 0	0.095 9	0.130 9	0.016 8
	模型权重	1.000 0	0.662 2	0.905 8	0.176 2
	平均权重	1.000 0	0.673 6	0.921 4	0.203 1
河流量	动态权重	0.867 6	0.068 5	0.093 4	0.760 0
	模型权重	0.904 1	0.135 5	0.185 2	1.549 6
	平均权重	0.908 7	0.148 0	0.202 2	1.738 5
风速	动态权重	0.870 1	0.128 1	0.174 8	0.762 6
	模型权重	0.976 1	0.314 1	0.429 7	1.261 1
	平均权重	0.976 3	0.339 2	0.463 9	1.528 7

数据集	多模态模型	整体模型		LSTM
数据集	多模态模型	RMSE	节点数	LSTM
MG	0.026 5	0.042 4	8	0.034 8
汇率	0.016 8	0.034 4	7	0.022 5
河流量	0.760 0	2.400 6	5	1.204 1
风速	0.762 6	0.802 8	10	0.841 6

数据集	多模态模型	整体模型	LSTM
MG	51.95	1.47	179.41
汇率	46.04	1.16	57.44
河流量	51.26	0.92	157.23
风速	335.05	4.06	4 137.44