Model-free forecasting for nonlinear time series (with application to exchange rates). (English) Zbl 0875.62432
Summary: The first step in nonlinear model construction is data identification. While the major drawback with most of the nonlinear tests is that they are designed to test a parametric null against certain parametric alternatives and are therefore inconsistent against all possibile alternatives. By incorporating the neurocomputing technique, the intriguing model construction and forecast problems such as how to find a best fitted model and to obtain a more appropriate prediction performance are addressed. The dynamic abilities of this approach lies in the fact that the multilier feed forward networks are functional approximators for any functions. Simulations verify that the neural networks perform a robust forecast for some nonlinear time series. Finally, we use the data of the exchange rates of Taiwan to compare the predictive performances of neural networks with ARIMA models.
MSC:
| 62M20 | Inference from stochastic processes and prediction |
| 62P20 | Applications of statistics to economics |
Keywords:
Model-free forecasting; Neural networks; Back-propagation method; Bilinear models; Exchange ratesReferences:
| [1] | Box, G. E. P.; Jenkins, G. M.: Time series analysis: forecasting and control. (1976) · Zbl 0363.62069 |
| [2] | Brockett, R. W.: Volterra series and geometric control theory. Automatica 12, 167-176 (1976) · Zbl 0342.93027 |
| [3] | Chan, W. S.; Tong, H.: On test for non-linearity in time series analysis. Intern. J. Forecasting 5, 217-228 (1986) |
| [4] | Cybento, G.: Approximation by superposition of a sigmoidal function. Mathematics of control, signals and systems 2, 303-314 (1989) · Zbl 0679.94019 |
| [5] | Degooijer, J. G.; Kumar, K.: Some recent developments in nonlinear time series modelling, testing and forecasting. Intern. J. Forecasting 8, 135-156 (1992) |
| [6] | Engle, R. F.: Autoregressive conditional heteroscedasticity with estimates of the variance of UK inflation. Econometrica 50, 987-1008 (1982) · Zbl 0491.62099 |
| [7] | Funahashi, K. I.: On the approximate of continuous mappings by neural networks. Neural networks 2, 183-192 (1989) |
| [8] | Granger, C. W. J.; Anderson, A. P.: An introduction to bilinear time series models. (1978) · Zbl 0379.62074 |
| [9] | Granger, C. W. J.: Developments in the nonlinear analysis of economic series. Scand. J. Of econ. 93, No. 2, 263-276 (1991) |
| [10] | Grossberg, S.: Studies of mind and brain: neural principles of learning, perception, development, cognition and motor control. (1933) · Zbl 0605.92017 |
| [11] | Guegan, D.; Pham, T. D.: Power of the score test against bilinear time series models. Statistica sinica 2, No. 1, 157-169 (1992) · Zbl 0820.62073 |
| [12] | Hecht-Nielsen, R.: Neurocomputing. IEEE spectrum, 36-41 (1989) |
| [13] | Hinich, M.: Testing for gaussianity and linearity of a stationary time series. J. time ser. Anal. 3, No. No. 3, 169-176 (1982) · Zbl 0502.62079 |
| [14] | Hornik, K.; Stinchcombe, M.; White, H.: Multilayer feedfoward networks are universal approximators. Neural networks 2, 359-366 (1989) · Zbl 1383.92015 |
| [15] | Kolen, J. F.; Goel, A. K.: Learning in parallel distributed processing networks: computational complexity and information content. IEEE transactions on systems, man, and cybernetics 21, No. 2, 359-367 (1991) |
| [16] | Kosko, B.: Neural networks for signal processing. (1992) · Zbl 0755.94024 |
| [17] | Lapedes, A.; Farber, R.: Nonlinear signal processing using neural networks: prediction and system modelling. Technical report LA-UR-87-2662 (1987) |
| [18] | Lapedes, A.; Farber, R.: How neural nets work. Theoretical division (1988) |
| [19] | Luukkonen, R.; Saikkonen, P.; Terasvirta, T.: Testing linearity against smooth transition autocorrelation models. Biometrica 75, 491-500 (1988) · Zbl 0657.62109 |
| [20] | Mckenzie, E.: Some simple models for discrete variate time series. Time series analysis in water resources, 645-650 (1935) |
| [21] | Pesaran, M. H.; Potter, S. M.: Nonlinear dynamics and econometrics: an introduction. J. appl. Econom. 7, S1-S7 (1992) |
| [22] | Priestley, M. B.: State-dependent models: A general approach to nonlinear time series. J. time series anal. 1, 47-71 (1980) · Zbl 0496.62076 |
| [23] | Rumelhart, D. E.; Hinton, G. E.; Williams, R. J.: Learning representations by back-propagation errors. Nature 323, 533-536 (1986) · Zbl 1369.68284 |
| [24] | Saikkonen, P.; Luukkonen, K.: Lagrange multiplier test for testing non-linearities in time series models. Scand. J. Of statist. 15, 55-68 (1988) · Zbl 0649.62087 |
| [25] | Rao, T. Subba; Gabr, M. M.: An introduction to bispectral analysis and bilinear time series models. Lecture notes in statistics (1984) · Zbl 0543.62074 |
| [26] | Tjøstheim, D.: Some doubly stochastic time series models. J. time ser. Anal. 7, 51-72 (1986) · Zbl 0588.62169 |
| [27] | Tong, H.; Lim, K. S.: Threshold autoregression, limit cycles and cyclical data. J. roy. Statist. soc. Ser. B 42, 245-292 (1980) · Zbl 0473.62081 |
| [28] | Tsay, R. S.: Detecting and modeling nonlinearity in univariate time series analysis. Statistica sinica 1, No. 2, 431-451 (1991) · Zbl 0824.62085 |
| [29] | Vogl, T. P.; Mangis, J. K.; Rigler, A. K.; Zink, W. T.; Alkon, D. L.: Accelerating the convergence of the back-propagation method. Biological cybernetics 59, 257-263 (1988) |
| [30] | Weiss, A. A.: ARCH and bilinear time series models: comparison and combination. J. business& econ. Statist. 4, No. No. 1, 59-70 (1986) |
| [31] | White, H.: Consequences and detection of misspecified nonlinaer regulartion models. J. amer. Statist. assoc. 76, 419-433 (1981) · Zbl 0467.62058 |
| [32] | Wu, B.; Liou, W.; Chen, Y.: Robust forecasting for the stochastic models and chaotic models. J. chinese statist. Assoc. 30, No. No. 2, 169-189 (1992) |
| [33] | Wu, B.; Shih, N.: On the identification problem for bilinear time series models. J. statist. Comput. simul. 43, 129-161 (1992) · Zbl 0836.62061 |
This reference list is based on information provided by the publisher or from digital mathematics libraries. Its items are heuristically matched to zbMATH identifiers and may contain data conversion errors. In some cases that data have been complemented/enhanced by data from zbMATH Open. This attempts to reflect the references listed in the original paper as accurately as possible without claiming completeness or a perfect matching.
