A New Increasing Translation Invariant Morphological Method for Financial Time Series Forecasting
Resumo
This paper presents a new method, referred to as Increasing Translation Invariant Morphological (ITIM), to overcome the random walk dilemma for financial time series forecasting. It consists of a hybrid intelligent model composed of a Modular Morphological Neural Network (MMNN) and a Modified Genetic Algorithm (MGA), which searches for the minimum number of time lags for a fine tuned time series representation, as well as by the initial weights, architecture and number of modules of the MMNN. Each element of the MGA population is trained via Back Propagation (BP) algorithm to further improve the parameters supplied by the MGA. The proposed method, after forecasting model adjustment, performs a behavioral statistical test and a phase fix procedure to adjust time phase distortions that appear in financial time series. An experimental analysis is conducted with the proposed method using two real world time series and five well-known performance measurements, demonstrating consistent better performance of this kind of morphological system.
Referências
Box, G. E. P., Jenkins, G. M., and Reinsel, G. C. (1994). Time Series Analysis: Forecasting and Control. Prentice Hall, New Jersey, third edition.
Clements, M. P., Franses, P. H., and Swanson, N. R. (2004). Forecasting economic and financial time-series with non-linear models. International Journal of Forecasting, 20(2):169–183.
Clements, M. P. and Hendry, D. F. (1993). On the limitations of comparing mean square forecast errors. Journal of Forecasting, 12(8):617–637.
Ferreira, T. A. E., Vasconcelos, G. C., and Adeodato, P. J. L. (2008). A new intelligent system methodology for time series forecasting with artificial neural networks. In Neural Processing Letters, volume 28, pages 113–129.
Hagan, M. and Menhaj, M. (1994). Training feedforward networks with the marquardt algorithm. IEEE Transactions on Neural Networks, 5(6):989–993.
Leung, F. H. F., Lam, H. K., Ling, S. H., and Tam, P. K. S. (2003). Tuning of the structure and parameters of the neural network using an improved genetic algorithm. IEEE Transactions on Neural Networks, 14(1):79–88.
Matheron, G. (1975). Random Sets and Integral Geometry. Wiley, New York.
Mills, T. C. (2003). The Econometric Modelling of Financial Time Series. Cambridge University Press, Cambridge.
Ozaki, T. (1985). Nonlinear Time Series Models and Dynamical Systems, volume 5 of Hand Book of Statistics. North-Holland, Amsterdam.
Pessoa, L. F. C. and Maragos, P. (1998). MRL-filters: A general class of nonlinear systems and their optimal design for image processing. IEEE Transactions on Image Processing, 7:966–978.
Prechelt, L. (1994). Proben1: A set of neural network benchmark problems and benchmarking rules. Technical Report 21/94.
Preminger, A. and Franck, R. (2007). Forecasting exchange rates: A robust regression approach. International Journal of Forecasting, 23(1):71–84.
Priestley, M. B. (1988). Non-Linear and Non-stationary Time Series Analysis. Academic Press.
Rao, T. S. and Gabr, M. M. (1984). Introduction to Bispectral Analysis and Bilinear Time Series Models, volume 24 of Lecture Notes in Statistics. Springer, Berlin.
Rumelhart, D. E. and McCleland, J. L. (1987). Parallel Distributed Processing, Explorations in the Microstructure of Cognition, volume 1 & 2. MIT Press.
Sousa, R. P. (2000). Design of translation invariant operators via neural network training. PhD thesis, UFPB, Campina Grande, Brazil.
Takens, F. (1980). Detecting strange attractor in turbulence. In Dold, A. and Eckmann, B., editors, Dynamical Systems and Turbulence, volume 898 of Lecture Notes in Mathematics, pages 366–381, New York. Springer-Verlag.
Zhang, G.P.; Kline, D. (Nov. 2007). Quarterly time-series forecasting with neural networks. Neural Networks, IEEE Transactions on, 18(6):1800–1814.
