Journal of Korean Institute of Industrial Engineers (대한산업공학회지)

Korean Institute of Industrial Engineers (대한산업공학회)
권호 표지

EWMA Based Fusion for Time Series Forecasting

시계열 예측을 위한 EWMA 퓨전

  • Shin, Hyung Won (Department of Computer Science & Industrial Systems Engineering 134, Yonsei University) ;
  • Sohn, So Young (Department of Computer Science & Industrial Systems Engineering 134, Yonsei University)
  • 신형원 (연세대학교 컴퓨터과학.산업시스템공학과) ;
  • 손소영 (연세대학교 컴퓨터과학.산업시스템공학과)
  • Published : 2002.06.30
Download PDF
⟨ Previous Next ⟩

Abstract

In this paper, we propose a new data fusion method to improve the performance of individual prediction models for time series data. Individual models used are ARIMA and neural network and their results are combined based on the weight reflecting the inverse of EWMA of squared prediction error of each individual model. Monte Carlo simulation is used to identify the situation where the proposed approach can take a vintage point over typical fusion methods which utilize MSE for weight. Study results indicate the following: EWMA performs better than MSE fusion when the data size is large with a relatively big amplitude, which is often observed in intra-cranial pressure data. Additionally, EWMA turns out to be a best choice among MSE fusion and the two individual prediction models when the data size is large with relatively small random noises, often appearing in tax revenue data.

Keywords

References

  1. Chiu, C. (1998), Prediction unemployment rates using fuzzy time series model and neural network, Proceedings of the lASTED International Conference Artificial Intelligence and Soft Computing, May 27-30, Cancun, Mexico
  2. Donaldson, R. G. and Kamstra. M. (1999), Neural network forecast combining with interaction effects, Journal of the Franklin Institute, 336, 227-236
  3. Figwer, J. (1997), A new Method of random time series simulation, Simulation Practice and Theory, 5, 217-234
  4. Hansen, J. V. and Nelson, R. D. (1997), Neural network and traditional time series methods: A synergistic combination in state economic forecasts, IEEE Transaction on Neural Networks, 8(4), 863-873 https://doi.org/10.1109/72.595884
  5. Ikonommopoulos, A. and Endou, A. (1998), Wavelet Decomposition and radial basis function networks for system monitoring IEEE Transactions on Nuclear science, 45(5), 2293-2301
  6. Ji, W-C. (1999), Decomposition analysis of time series using neural networks. Journal of the Korean Institute of Industrial Engineers, 25(1), 111-114
  7. Khotanzad, A., Elragal, H. and Lu, T. (2000), Combination of artificial neural-network forecasts for prediction of natural gas consumption, IEEE Transaction on Neural Networks, 11(2), 464-473
  8. Kiartzis, S., Kehagias, A. and Bakirtzis, A. (1997), Short term load forecasting using a bayesian combination method, Electrical Power & Energy Systems, 19(3), 171-177 https://doi.org/10.1016/S0142-0615(96)00038-5
  9. Kim, D. (1998), Improving the fuzzy system performance by fuzzy system ensemble, Fuzzy Set and Systems, 98, 43-56
  10. Kim, I-T. and Gong, Cow. (1997), Fuzzy learning algorithms for time series prediction, Journal of the Korean Institute of Fuzzy & Intelligence Systems, 7(3), 34-42
  11. Kohzadi, N., Boyd, M. S., Kermanshahi, B. and Kaastra, L. (1996), A comparison of artificial neural network and time series model for forecasting commodity prices, Neuro Computing, 10, 169-181
  12. Mignani, R. and Rosa, R. (1995), The moving block boostrap to assess the accuracy of statistical estimates in Ising model simulation, Computer Physics Communications, 92, 203-213
  13. Mol, R. M., Keen, A., Kroeze, G. H. and Achten, J. M. F. H. (1999), Description of a detection model for oestrus and disease in dairy cattle based on time series analysis combined with a kalman filter, Computers and Electronics in Agriculture, 22, 171-185
  14. Montgomery, D. C. (1996a), Introduction to Statistical Quality Control, (3th ed), John Wiley & Sons
  15. Montgomery, D. C. (1996b), Design and analysis of experiments, (3th ed). John Wiley & Sons
  16. Ridley, D. (1995), Combining global Antithetic Forecasts, International Transaction in Operational Research, 2(4), 387-398 https://doi.org/10.1111/j.1475-3995.1995.tb00030.x
  17. Ridley, D. (1999), Optimal antithetic weights for lognormal time series forecasting, Computer & Operations Research, 26, 189-209
  18. Roberts (1959), Control chart tests based on geometric moving average, Technometric, 1
  19. Schwaerzel, R. and Rosen, B. (1997), Improving the Accuracy of financial time Series Prediction Using Ensemble Networks and High Order Statistics, Proceedings of the 1997 International Conference on Neural Networks(lCNN'97), 4, 2045-2050
  20. Shi, S. M., Xu, L. D. and Boo, L. (1999), Improving the accuracy of linear combined forecasting using neural networks, Expert Systems with Applications, 16, 49-54
  21. Swiercz, M., Mariak, Z., Lewko, J., Chojnacki, K., Kozlowski, A. and Piekarski, P. (1998), Neural Network technique for detecting emergency states in neurosurgical patients, Medical & Biological Engineering & Computing, 36, 717-722
  22. Trygg, J. and Wold, S. (1998), PLS regression on wavelet compressed NIR spectra, Chemometrics and Intelligent Laboratory Systems, 42, 209-220
  23. Zhang, B. and Joung, J. (1999), Time Series Prediction Using Committee Machines of Evolutionary Neural Trees, Proceedings of the 1999 Congress on Evolutionary Computation, 1, 281-286