An analysis of adaptive windowing for time series forecasting in dynamic environments: further tests of the DyFor GP model

ABSTRACT

Genetic Programming (GP) has proved its applicability for time series forecasting in a number of studies. The Dynamic Forecasting Genetic Program (DyFor GP) model builds on the GP technique by adding features that are tailored for the forecasting of time series whose underlying data-generating processes are non-static. Such time series often appear for real-world forecasting concerns in which environmental conditions are constantly changing. In a previous study the DyFor GP model was shown to improve upon the performance of GP and other benchmark models for a set of simulated and real time series. The distinctive feature of DyFor GP is its adaptive data window adjustment. This feedback-driven window adjustment is designed to automatically hone in on the currently active process in an environment where the generating process varies over time. This study further investigates this adaptive windowing technique and provides an analysis of its dynamics for constructed time series with non-static data-generating processes. Results show that DyFor GP is able to capture the moving processes more accurately than standard GP and offer insight for further improvements to DyFor GP.

REFERENCES

• 1: Eiben A., Hinterding R., and Michalewicz Z. Parameter control in evolutionary algorithms. IEEE Transactions on Evolutionary Computation, 3:124--141, 1999.
• 2: Henon M. A two-dimensional mapping with a strange attractor. Communications in Mathematical Physics, 50:69, 1976.
• 3: Iba H. and Nikolaev N. Genetic programming polynomial models of financial data series. In Proceedings of the 2000 Congress of Evolutionary Computation, pages 1459--1466. IEEE, 2000.
• 4: Kaboudan M. Genetic evolution of regression models for business and economic forecasting. In Proceedings of the 1999 Congress of Evolutionary Computation, pages 1260--1268. IEEE, 1999.
• 5: M. A. Kaboudan, Genetic Programming Prediction of Stock Prices, Computational Economics, v.16 n.3, p.207-236, Dec. 2000 [doi>10.1023/A:1008768404046]
• 6: Kaboudan M. Genetically evolved models and normality of their residuals. Journal of Economics Dynamics and Control, 25:1719--1749, 2001.
• 7: Kaboudan M. Forecasting with computer-evolved model specifications: a genetic programming application. Computer and Operations Research, 30:1661--1681, 2003.
• 8: John R. Koza, Genetic programming: on the programming of computers by means of natural selection, MIT Press, Cambridge, MA, 1992
• 9: Mackey M. and Glass L. Oscillation and chaos in physiological control systems. Science, 197:287, 1977.
• 10: Makridakis, S., Wheelwright, S., and Hyndman, R. Forecasting: Methods and Applications. John Wiley and Sons, Inc., Hoboken, NJ, 1998.
• 11: May R. Simple mathematical models with very complicated dynamics. Nature, 261:459--467, 1976.
• 12: May R. Simple mathematical models with very complicated dynamics. Nature, 261:459--467, 1976.
• 13: Neely C. and Weller P. Predicting exchange rate volatility: genetic programming versus GARCH and RiskMetrics. Technical report, The Federal Reserve Bank of St. Louis, 2002.
• 14: Oakley H. Genetic programming for nonlinear equation fitting to chaotic data. In Back T., Fogel D., and Michalewicz Z., editor, Handbook of Evolutionary Computation. Oxford University Press, 1997.
• 15: Ozaki, T. Non-linear time series models and dynamical systems. In Hannan E., Krishnaiah P., Rao M., editor, Handbook of Statistics. Elsevier, 1985.
• 16: South Australian Partnership for Advanced Computing. High performance and grid computing - hydra cluster. http://www.sapac.edu.au/systems/hydra/index.html, 2007.
• 17: Wagner N. and Michalewicz Z. Genetic programming with efficient population control for financial times series prediction. In 2001 Genetic and Evolutionary Computation Conference Late Breaking Papers, pages 458--462, 2001.
• 18: Wagner N., Michalewicz Z., Khouja M., and McGregor R. Time series forecasting for dynamic environments: the DyFor genetic program model. IEEE Transactions on Evolutionary Computation, 11(4):433--452, 2007.

INDEX TERMS

Primary Classification:
  I. Computing Methodologies
  I.2 ARTIFICIAL INTELLIGENCE
      I.2.8 Problem Solving, Control Methods, and Search
          Subjects: Heuristic methods

Additional Classification:
  I. Computing Methodologies
  I.6 SIMULATION AND MODELING
      I.6.5 Model Development

General Terms:
Algorithms

Keywords:
dynamic, forecasting, genetic programming, time series, uncertain environments