Local dynamic integration of ensemble in prediction of time series

• Autorzy: Osowski S. , Siwek K.

Identyfikatory

DOI

10.24425/bpasts.2019.129650

• Języki publikacji: EN

Abstrakty

EN

The paper presents local dynamic approach to integration of an ensemble of predictors. The classical fusing of many predictor results takes into account all units and takes the weighted average of the results of all units forming the ensemble. This paper proposes different approach. The prediction of time series for the next day is done here by only one member of an ensemble, which was the best in the learning stage for the input vector, closest to the input data actually applied. Thanks to such arrangement we avoid the situation in which the worst unit reduces the accuracy of the whole ensemble. This way we obtain an increased level of statistical forecasting accuracy, since each task is performed by the best suited predictor. Moreover, such arrangement of integration allows for using units of very different quality without decreasing the quality of final prediction. The numerical experiments performed for forecasting the next input, the average PM10 pollution and forecasting the 24-element vector of hourly load of the power system have confirmed the superiority of the presented approach. All quality measures of forecast have been significantly improved.

Słowa kluczowe

EN

neural networks; ensemble of predictors; dynamic integration; time series prediction

PL

sieci neuronowe; zespół predyktorów; integracja dynamiczna

• Wydawca: Polska Akademia Nauk, Wydział IV Nauk Technicznych
• Czasopismo: Bulletin of the Polish Academy of Sciences. Technical Sciences
• Rocznik: 2019
• Tom: Vol. 67, nr 3
• Strony: 517--525
• Opis fizyczny: Bibliogr. 26 poz., wykr., tab.

Twórcy

autor

Osowski S.

• Military University of Technology, Warsaw, Poland

autor

Siwek K.

• Warsaw University of Technology, Warsaw, Poland

Bibliografia

• [1] H. Taheri Shahraiyni and S. Sodoudi, “Statistical modeling approaches for PM10 prediction in urban areas; A review of 21st-century studies”, Atmosphere, 7, 1?24, doi:10.3390/atmos7020015 (2016).
• [2] A. Paschalidou, S. Karakitsios, S. Kleanthous, and P. Kassomenos, “Forecasting hourly PM10 concentration in Cyprus through artificial neural networks and multiple regression models: implications to local environmental management”, Environ. Sci. Pollut. Res. DOI 10.1007/s11356?010?0375?2 (2010).
• [3] A. Z. Ul-Saufie, S. Yahya, and N. A Ramli, “Comparison between multiple linear regression and feed forward back propagation neural network models for predicting PM10 concentration level based on gaseous and meteorological parameters”, Intern. J. of Applied Science and Technology 1 (4), 42?49 (2011).
• [4] K. Siwek and S. Osowski, “Data mining methods for prediction of air pollution”, International Journal of Applied Mathematics and Computer Science 26 (2), 467–478 (2016).
• [5] G. Gennaro, L. Trizio, A. Gilio, J. Pey, N. Pérez, M. Cusack, A. Alastuey, and X. Querol, “Neural network model for the prediction of PM10 daily concentrations in two sites in the Western Mediterranean”, Science of The Total Environment 463?464, 875–883 (2013).
• [6] A. Daly and P. Zannetti, “Air Pollution Modeling – An Overview”, in chapter 2 of “Ambient air pollution” of P. Zannetti, D. Al-Ajmi, and S. Al-Rashied (eds). The EnviroComp Institute, 15?28 http://www.envirocomp.org/ (2007).
• [7] M. Marko, E. Kyriakides, and M. Polycarpou, “24-Hour ahead short term load forecasting using multiple MLP”, Intern. Conference on Deregulated Electricity Market Issues in South-Eastern Europe (DEMSEE), Cyprus (2008).
• [8] S. Shah, H.N. Nagaraja, and J. Chakravorty, “Short term load forecasting model for UGVCL, MGVCL, DGVCL and PGVCL using ANN”, International Journal of Recent Trends in Electrical & Electronics Eng. 5(2):21?30 (2017).
• [9] D. Niu, Y. Wang, and D.D. Wu, “Power load forecasting using SVM and ant colony optimization”, Expert Systems with Applications 37:2531?2539 (2010).
• [10] R. Chandra and M. Zhang, “Cooperative coevolution of Elman recurrent neural networks for chaotic time series prediction”, Neurocomputing 86, 116?123 (2012).
• [11] M. Luzar, Ł. Sobolewski, W. Miczulski, and J. Korbicz, “Prediction of corrections for the Polish time scale UTC(PL) using artificial neural networks”, Bull. Pol. Ac.: Tech. 61(3), 589?594 (2013).
• [12] M. Walenczykowska and A. Kawalec, “Type of modulation identification using Wavelet Transform and Neural Network”, Bull. Pol. Ac.: Tech. 64(1), 257?261 (2016).
• [13] S. Osowski, K. Siwek, and R. Szupiluk, “Ensemble neural network approach for accurate load forecasting in the power system”, International Journal of Applied Mathematics and Computer Science, 19 (2), 303?315 (2009).
• [14] L. Kuncheva, Combining Pattern Classifiers: Methods and Algorithms, Wiley, New York ( 2004).
• [15] L. Xu, A. Krzyzak, and C.Y. Suen, “Methods of combining multiple classifiers and their applications to handwriting recognition”, IEEE Trans. Systems, Man, and Cybernetics 22 (3), 418–435 (1992).
• [16] A.S. Britto, R. Sabourin, and L. Oliveira, “Dynamic selection of classifiers – a comprehensive review”, Pattern Recognition 47, 3665–3680 (2014).
• [17] K. Woods, W.P. Kegelmeyer, and K. Bowyer, “Combination of multiple classifiers using local accuracy estimates”, IEEE Trans. Pattern Analysis and Machine Intelligence 19 (4), 405?410 (1997).
• [18] M. Martuzzi, F. Mitis, I. Iavarone, and M. Serinelli, “Health impact of PM10 and ozone in 13 Italian cities” WHP report (2005).
• [19] L. Nikias and A. Petropulu, Higher order spectral analysis, Prentice Hall, New York (1993).
• [20] Matlab user manual, MathWorks, Natick, USA, (2016).
• [21] I. Guyon and A. Elisseeff, “An introduction to variable and feature selection”, J. Mach. Learn. Res. 3, 1157?1182 (2003).
• [22] R.L. Haupt and S.E. Haupt, Practical Genetic Algorithms, Wiley, New York (2004).
• [23] B. Schölkopf and A. Smola, Learning with kernels, MIT Press, Cambridge MA.,USA, 2002.
• [24] I. Goodfellow, Y. Bengio, and A. Courville, 2016. Deep learning, MIT Press, Cambridge (2016).
• [25] P. Romeu, F. Zamora-Martinez, P. Botella-Rocamora, and J. Pardo, “Stacked denoising autoencoders for short-term time series forecasting”, in P. Koprinkova-Hristova et al. (eds), Artificial Neural Networks, Springer series in Bio-Neuroinformatics 4, 2015, doi: 10.1007/978?3-319?09903?3_23 (2015).
• [26] F. Taşpinar, “Improving artificial neural network model predictions of daily average PM10 concentrations by applying principle component analysis and implementing seasonal model”, Journal of the Air & Waste Management Association 65, 800?809 (2015).

Uwagi

PL

Opracowanie rekordu w ramach umowy 509/P-DUN/2018 ze środków MNiSW przeznaczonych na działalność upowszechniającą naukę (2019).