Please wait a minute...
Frontiers of Computer Science

ISSN 2095-2228

ISSN 2095-2236(Online)

CN 10-1014/TP

Postal Subscription Code 80-970

2018 Impact Factor: 1.129

Front Comput Sci Chin    2010, Vol. 4 Issue (2) : 254-262    https://doi.org/10.1007/s11704-010-0506-4
RESEARCH ARTICLE
Neural network methods for forecasting turning points in economic time series: an asymmetric verification to business cycles
Dabin ZHANG1,2, Lean YU2(), Shouyang WANG2, Haibin XIE2
1. Department of Information Management, Huazhong Normal University, Wuhan 430079, China; 2. Institute of Systems Science, Academy of Mathematics and Systems Science, Chinese Academy of Sciences, Beijing 100190, China
 Download: PDF(233 KB)   HTML
 Export: BibTeX | EndNote | Reference Manager | ProCite | RefWorks
Abstract

This paper examines the relevance of various financial and economic indicators in forecasting business cycle turning points using neural network (NN) models. A three-layer feed-forward neural network model is used to forecast turning points in the business cycle of China. The NN model uses 13 indicators of economic activity as inputs and produces the probability of a recession as its output. Different indicators are ranked in terms of their effectiveness of predicting recessions in China. Out-of-sample results show that some financial and economic indicators, such as steel output, M2, Pig iron yield, and the freight volume of the entire society are useful for predicting recession in China using neural networks. The asymmetry of business cycle can be verified using our NN method.

Keywords turning points      business cycle      leading indicators      neural networks (NNs)     
Corresponding Author(s): YU Lean,Email:yulean@amss.ac.cn   
Issue Date: 05 June 2010
 Cite this article:   
Dabin ZHANG,Shouyang WANG,Haibin XIE, et al. Neural network methods for forecasting turning points in economic time series: an asymmetric verification to business cycles[J]. Front Comput Sci Chin, 2010, 4(2): 254-262.
 URL:  
https://academic.hep.com.cn/fcs/EN/10.1007/s11704-010-0506-4
https://academic.hep.com.cn/fcs/EN/Y2010/V4/I2/254
Fig.1  A three-layer feed-forward neural network, (assuming >2)
Fig.1  A three-layer feed-forward neural network, (assuming >2)
IndicatorCodeDescriptionPublishing mechanism
PIY1pig iron yieldNIC
SO2steel outputNIC
CTCP3cargo throughput of coastal portNIC, NSB
NSOACH4new start operation area of commercial housingNIC, NSB
FEB5financial expenditure budgetNIC
COVFG6capital occupying volume of finished goods(reversal)NIC
M27M2NSB
SRIP8sales rate of industrial productsNSB
HSI9Hang Seng IndexNSB
NDAL10new developed area of landNSB
CI11consumer indexNSB
PPNSO12plan projects of new start operationNSB
FVWS13freight volume of whole societyNSB
Tab.1  Economic indicators
Fig.2  Peaks and troughs of industrial added value (P: peak; T: tough)
Fig.2  Peaks and troughs of industrial added value (P: peak; T: tough)
Rankgh = 2h = 3h = 4h = 5h = 6h = 7h = 8
iMSFEiMSFEiMSFEiMSFEiMSFEiMSFEiMSFEiMSFE
120.14120.13920.15720.15020.18320.20420.21320.200
210.141110.16270.191110.20970.23970.258130.252130.263
3110.14210.170110.19870.216130.24480.26280.26760.283
480.16470.17860.20310.22110.245110.26770.27680.291
560.16580.183130.21980.232110.249130.270110.283110.293
6130.17260.18810.219130.23680.25460.28560.29110.296
770.178130.19630.22940.23760.25710.28810.31270.327
830.19030.20080.22930.24030.27830.29940.32940.329
990.19340.22240.23260.24240.28540.31730.33730.355
1040.19390.237120.237120.308100.34290.360100.38890.383
11120.209100.24990.28590.311120.350100.38190.394100.406
1250.214120.255100.309100.33590.352120.394120.412120.414
13100.23650.26250.30950.38850.43550.47150.49250.487
Tab.2  Out-of-sample forecasts, and rank of economic indicators ( = 1-8)
Fig.3  Composite index of different indicators. (a) Composite index of NIC; (b) composite index of NSB; (c) composite index of all indicators; (d) composite index of optimization indicators
Fig.3  Composite index of different indicators. (a) Composite index of NIC; (b) composite index of NSB; (c) composite index of all indicators; (d) composite index of optimization indicators
Rankh = 1h = 2h = 3h = 4h = 5h = 6h = 7h = 8
iMSFEiMSFEiMSFEiMSFEiMSFEiMSFEiMSFEiMSFE
1110.138110.147110.147110.114110.114110.114110.106110.106
2130.163130.171130.179130.17160.201100.23730.16930.206
350.19180.18080.180100.212130.21260.237100.26220.229
4120.19230.18260.18380.22130.21430.23820.262100.262
560.19360.199100.23730.232100.224130.253130.27760.290
610.19950.19930.24050.23220.24520.26260.30140.291
7100.212100.21250.24160.23780.246120.290120.315120.307
830.21270.21320.27720.24550.26650.29140.31650.307
980.221120.24770.288120.248120.26610.29950.316130.310
1020.22810.249120.29070.27170.29640.30710.33210.340
1170.23020.27810.29110.27410.30780.32080.35370.386
1240.25740.28240.29140.29140.32470.32070.36190.386
1390.28790.38590.38690.42790.43590.46890.44380.386
Tab.3  Out-of-sample and rankings of individual indicators, 1-8 months ahead, in first sub-period (January 1998 to December 2002)
Rankh = 1h = 2h = 3h = 4h = 5h = 6h = 7h = 8
iMSFEiMSFEiMSFEiMSFEiMSFEiMSFEiMSFEiMSFE
180.129110.169110.16180.17780.23480.24290.25070.185
2110.14580.20120.193110.201110.258110.25070.25880.201
360.16990.20180.21050.22650.25890.27480.26620.258
4130.18560.21070.21870.24270.27470.28220.26690.266
590.19370.21890.22690.24220.29020.298110.322110.290
610.21010.22650.24220.26690.30650.33830.33050.314
720.210130.23410.25010.27410.33860.33850.33830.314
870.21020.242100.26660.290100.355100.355100.37910.355
950.21850.25860.282100.29060.35530.35510.387100.371
1030.21830.266130.29030.322120.35510.37140.39560.371
11100.258100.27430.290130.33830.379120.38760.411120.419
1240.266120.31440.306120.35540.40340.427120.44340.427
13120.26640.330120.35540.363130.427130.467130.492130.500
Tab.4  Out-of-sample and rankings of individual indicators, 1-8 months ahead, in second sub-period (January 2003 to December 2008)
1 Auerbach A J. The index of leading indicators: “measurement without theory,” thirty-five years later. Review of Economics and Statistics , 1982, 64(4): 589-595
doi: 10.2307/1923943
2 Kaufmann S. Measuring business cycle with a dynamic Markov switching factor model: An assessment using Bayesian simulation methods. Econometrics Journal , 2000, 3(1): 39-65
doi: 10.1111/1368-423X.00038
3 Robert I, Jan J, Ward R. Business cycle indexes: Does a heap of data help? Journal of Business Cycle Measurement and Analysis , 2004, 1(3): 309-336
4 George E N, Ghassan D, Antoine A. Predicting business cycle turning points with neural network in an information-poor economy. In: Proceedings of The 2007 summer computer simulation conference (SCSC 2007) , 2007: 627-631
5 Zhang D B, Yu L, Wang S Y, Song Y W. A novel PPGA-based clustering analysis method for business cycle indicator selection. Frontiers of Computer Science in China , 2009, 3(2): 217-225
doi: 10.1007/s11704-009-0023-5
6 Neftci S N. Are economic time series asymmetric over the business cycle? Journal of Political Economy , 1984, 92(2): 307-328
doi: 10.1086/261226
7 Sichel D E. Are business cycles asymmetric? A correction. Journal of Political Economy , 1989, 97(5): 1255-1260
doi: 10.1086/261652
8 Quandt R E. A new approach to estimating switching regressions. Journal of the American Statistical Association , 1972, 67(338): 306-310
doi: 10.2307/2284373
9 Goldfeld S M, Quandt R E. A Markov model for switching regression. Journal of Econometrics , 1973, 1(1): 3-15
doi: 10.1016/0304-4076(73)90002-X
10 Ploberger W, Kr?mer W, Kontrus K. A new test for structural stability in the linear regression model. Journal of Econometrics , 1989, 40(2): 307-318
doi: 10.1016/0304-4076(89)90087-0
11 Wang J M, Gao T M, McNown R. Measuring Chinese business cycle with dynamic factor models. Journal of Asian Economics , 2009, 20(2): 89-97
doi: 10.1016/j.asieco.2008.10.003
12 Diebold F X, Rudebusch G D. Measuring business cycles: A modern perspective. Review of Economics and Statistics , 1996, 78(1): 67-77
doi: 10.2307/2109848
13 Chauvet M. An econometric characterization of business cycle dynamics with factor structure and regime switching. International Economic Review , 1998, 39(4): 969-996
doi: 10.2307/2527348
14 Kim C J, Nelson C R. Business cycle turning points, a new coincident index, and tests of duration dependence based on a dynamic factor model with regime switching. Review of Economics and Statistics , 1998, 80(2): 188-201
doi: 10.1162/003465398557447
15 Hoptroff R G, Bramson M J, Hall T J. Forecasting economic turning points with neural nets. In: Proceedings of the 1991 IEEE International Joint Conference on Neural Networks . 1991: 347-352
16 Vishwakarma K P. A neural network for signal modeling in business cycle studies. In: Proceedings of 1994 IEEE International Conference on Systems, Man, and Cybernetics, ‘Humans, Information and Technologyapos’ , 1994, 10: 2437-2442
17 Vishwakarma K P. Recognizing business cycle turning points by means of a neural network. Computational Economics , 1994, 7(3): 175-185
doi: 10.1007/BF01299778
18 Soo H C, Joon S L. Economic turning point forecasting using neural network with weighted fuzzy membership functions. Lecture Notes in Computer Science , 2007, 4570: 145-154
doi: 10.1007/978-3-540-73325-6_15
19 Qi M. Predicting US recessions with leading indicators via neural network models. International Journal of Forecasting , 2001, 17(3): 383-401
doi: 10.1016/S0169-2070(01)00092-9
20 Inoue A, Kilian L. In-Sample or Out-of-Sample Tests of Predictability: Which One Should We Use?ECB Working Paper , 2002, 11No. 195
21 Jagielska I, Jaworshi J. Neural networks for predicting the performance of credit card accounts. Computational Economics , 1996, 9(1): 77-82
doi: 10.1007/BF00115693
22 Romero R D, Touretzky D S, Thibadeau R H. Optical Chinese character recognition using probabilistic neural networks. Pattern Recognition , 1997, 30(8): 1279-1292
doi: 10.1016/S0031-3203(96)00166-5
23 Uncini A. Audio signal processing by neural networks. Neurocomputing , 2003, 55(3-4): 593-625
doi: 10.1016/S0925-2312(03)00395-3
24 Kondo T. Evolutionary design and behavior analysis of neuromodulatory neural networks for mobile robots control. Applied Soft Computing , 2007, 7(1): 189-202
doi: 10.1016/j.asoc.2005.05.004
25 Bailey L D, Donna T. How to develop neural network applications. AI Expert , 1990, 5(6): 38-47
26 Bailey L D, Donna T. Developing neural network applications. AI Expert , 1990, 5(9): 34-41
27 Tamura S. Capabilities of a three layer feed-forward neural network. 1991 IEEE International Joint Conference on Neural Networks , 1991, 11: 2757-2762
28 Hamilton J D, Perez-Quiros G. What do the leading indicators lead? Journal of Business , 1996, 69(1): 27-49
doi: 10.1086/209678
29 http://www.cemac.org.cn/indexbci.htm
30 Farley A M, Jones S. Using a genetic algorithm to determine an index of leading economic indicators. Computational Economics , 1994, 7(3): 163-173
doi: 10.1007/BF01299777
31 Layton A P, Moore G H. Leading indicators for the service sector. Journal of Business & Economic Statistics , 1989, 7(3): 379-386
doi: 10.2307/1391534
32 Stock J. H. and Watson M. W. New indexes of coincident and leading economic indicators. NBER Macroeconomics Annual 1989 , 1989: 351-294
33 Banerji A, Hiris L. A framework for measuring international business cycles. International Journal of Forecasting , 2001, 17(3): 333-348
doi: 10.1016/S0169-2070(01)00089-9
34 Zhang Y J. Research on econometric methods and application of business cycle. China Economic Publishing House , 2007, 11: 73-87
35 Layton A P. Dating and predicting phase changes in the U.S. business cycle. International Journal of Forecasting , 1996, 12(3): 417-428
doi: 10.1016/0169-2070(95)00663-X
[1] Dabin ZHANG, Lean YU, Shouyang WANG, Yingwen SONG. A novel PPGA-based clustering analysis method for business cycle indicator selection[J]. Front Comput Sci Chin, 2009, 3(2): 217-225.
Viewed
Full text


Abstract

Cited

  Shared   
  Discussed