Independent Component Analysis for Ensemble Predictors with Small Number of Models

• Authors: R. Szupiluk , T. Ząbkowski , K. Gajowniczek
• Languages of publication: EN

Abstracts

EN

The article presents independent component analysis (ICA) applied to the concept of ensemble predictors. The use of ICA decomposition enables to extract components with particular statistical properties that can be interpreted as destructive or constructive for the prediction. Such process can be treated as noise filtration from multivariate observation data, in which observed data consist prediction results. As a consequence of the ICA multivariate approach, the final results are combination of the primary models, what can be interpreted as aggregation step. The key issue of the presented method is the identification of noise components. For this purpose, a new method for evaluating the randomness of the signals was developed. The experimental results show that presented approach is effective for ensemble prediction taking into account different prediction criteria and even small set of models.

Keywords

EN

05.45.Tp; 05.40.Ca; 07.05.Kf; 07.05.Mh

Discipline

• 05.40.Ca: Noise
• 05.45.Tp: Time series analysis
• 07.05.Mh: Neural networks, fuzzy logic, artificial intelligence
• 07.05.Kf: Data analysis: algorithms and implementation; data management(for data analysis in nuclear physics, see 29.85.-c)

• Publisher: Institute of Physics, Polish Academy of Sciences
• Journal: Acta Physica Polonica A
• Year: 2015
• Volume: 127
• Issue: 3A
• Pages: A-139-A-144

Dates

published

2015-03

Contributors

author

R. Szupiluk

• Warsaw School of Economics, Niepodleglosci 162, 02-554 Warsaw,

author

T. Ząbkowski

• Warsaw University of Life Sciences, Faculty of Applied Informatics and Mathematics, Nowoursynowska 159, 02-776 Warsaw, Poland

author

K. Gajowniczek

• Warsaw University of Life Sciences, Faculty of Applied Informatics and Mathematics, Nowoursynowska 159, 02-776 Warsaw, Poland
• Systems Research Institute, Polish Academy of Sciences, Newelska 6, 01-447 Warsaw, Poland

References

• [1] R. Szupiluk, P. Wojewnik, T. Ząbkowski, Lect. Notes Comput. Sc. 6594, 206 (2011), doi: 10.1007/978-3-642-20267-4_22
• [2] P. Comon, Ch. Jutten, Handbook of Blind Source Separation: Independent Component Analysis and Applications, Academic Press, 2010
• [3] L. Breiman, Mach. Learn. 24, 123 (1996), doi: 10.1007/BF00058655
• [4] R.T. Clemen, Int. J. Forecasting 5, 559 (1989), doi: 10.1016/0169-2070(89)90012-5
• [5] J.H. Friedman, Ann. Stat. 19, 1 (1991), doi: 10.1214/aos/1176347963
• [6] A. Timmermann, Handbook of Economic Forecasting 1, 135 (2006), doi: 10.1016/S1574-0706(05)01004-9
• [7] J.W. Taylor, L.M. de Menezes, P.E. McSharry, Int. J. Forecasting 22, 1 (2006), doi: 10.1016/j.ijforecast.2005.06.006
• [8] R. Weron, Modeling and Forecasting Electricity Loads and Prices: A Statistical Approach, John Wiley, Chichester 2006
• [9] A. Cichocki, JCMSI 7, 507 (2013)
• [10] G. Zhou, A. Cichocki, Q. Zhao, S. Xie, IEEE Signal Proc. Mag. 31, 54 (2014), doi: 10.1109/MSP.2014.2298891
• [11] R. Szupiluk, P. Wojewnik, T. Ząbkowski, IJICIC 10, 1435 (2014)
• [12] R. Szupiluk, Przegląd Elektrotechniczny (Electrotechnical Review) 86, 144 (2010)
• [13] R. Szupiluk, P. Wojewnik, T. Zabkowski, Lect. Notes Artif. Int. 6113, 471 (2010), doi: 10.1007/978-3-642-13208-7_59
• [14] H.E. Hurst, Trans. Am. Soc. Civil Engineers 116, 770 (1951)
• [15] E. Peters, Fractal market analysis, John Wiley, Chichester 1996
• [16] A. Cichocki, S. Amari, Adaptive Blind Signal and Image Processing: Learning Algorithms and Applications, John Wiley, Chichester 2003

Identifiers

DOI

10.12693/APhysPolA.127.A-139