A novel hybrid data-driven approach is developed for forecasting power system parameters with the goal of increasing the efficiency of short-term forecasting studies for non-stationary time-series. The proposed approach is based on mode decomposition and a feature analysis of initial retrospective data using the Hilbert-Huang transform and machine learning algorithms. The random forests and gradient boosting trees learning techniques were examined. The decision tree techniques were used to rank the importance of variables employed in the forecasting models. The Mean Decrease Gini index is employed as an impurity function. The resulting hybrid forecasting models employ the radial basis function neural network and support vector regression.

A part from introduction and references the paper is organized as follows. The second section presents the background and the review of several approaches for short-term forecasting of power system parameters. In the third section a hybrid machine learningbased algorithm using Hilbert-Huang transform is developed for short-term forecasting of power system parameters. Fourth section describes the decision tree learning algorithms used for the issue of variables importance. Finally in section six the experimental results in the following electric power problems are presented: active power flow forecasting, electricity price forecasting and for the wind speed and direction forecasting.

1. Areekul Ph. Senjyu T., Toyama H., Yona A. A Hybrid ARIMA and Neural Network Model for Short-Term Price Forecasting in Deregulated Market. IEEE Trans. Power Syst, 2010, vol. 25, no 1, pp. 524-530.

2. Breiman L., Friedman J., Olshen R., Stone C. Classification and Regression Trees. Belmont California, Wadsworth, 1984.

3. Breiman L. Random Forests. Machine Learning, 2001, vol. 45, pp. 5-32.

4. Damousis I.G., Dokopoulos P. A fuzzy model expert system for the forecasting of wind speed and power generation in wind farms. Proc. of the IEEE International Conference on Power Industry Computer Applications PICA 01, 2001, pp. 63-69.

5. Friedman J. Greedy Function Approximation: A Gradient Boosting Machine. Annals of Statistics, 2001, vol. 29.

6. Foley A.M. Leahy P.G., Marvuglia A., McKeogh E.J. Current Methods and Advances in Forecasting of Wind Power Generation. Renewable Energy, 2012, vol. 37, no 1, pp. 1-8.

7. Friedman J.H. Stochastic gradient boosting. Computational Statistics and Data Analysis, 2002, vol. 38, pp. 367-378.

8. Fugon L., Juban J., Kariniotakis G. Data mining for wind power forecasting. Proc. the European Wind Energy Conference, Brussels, Belgium, Apr. 2008.

9. Garcia R.C., Contreras J., M. van Akkeren, Batista J., Garcia C. A GARCH Forecasting Model to Predict Day-Ahead Electricity Prices. IEEE Transactions on Power Systems, 2005, vol. 20, no 2.

10. Gerikh V., Kolosok I., Kurbatsky V., Tomin N. Application of Neural Network Technologies for Price Forecasting in the Liberalized Electricity Market. Scien. Journal of Riga Tech. Univer. Ser. Power and Electrical Engineering, 2009, vol. 5, pp. 91-96.

11. Glazunova A.M., Forecasting Power System State Variables on the Basis of Dynamic State Estimation and Artificial Neural Networks. Proc. the IEEE Region 8 SIBIRCON-2010, Listvyanka, Russia, 2010

12. Hippert H. Steinherz, Pedreira C. Eduardo, Souza R. Castro. Neural Networks for Short-Term Load Forecasting: A Review and Evaluation. IEEE Trans. On Power Systems, 2001, vol. 16, no 1.

13. Kavasseri R.G., Seetharaman K. Day-ahead wind speed forecasting using f-ARIMA models. Renewable Energy, 2009, vol. 34, no. 5, pp. 1388-1393.

14. Khotanzad A.Z., Elragal H. A neuro-fuzzy approach to short-term load forecasting in a price sensitive environment. IEEE Trans. Power Syst, 2002, vol. 17, no 4, pp. l273-1282.

15. Kurbatsky V. Sidorov D., Spiryaev V., Tomin N. On the Neural Network Approach for Forecasting of Nonstationary Time Series on the Basis of the Hilbert–Huang Transform. Automation and Remote Control, 2011, vol. 72, no. 7, pp. 1405-1414.

16. Kurbatsky V., Tomin N., Sidorov D., Spiryaev V. Hybrid Model for Short-Term Forecasting in Electric Power System. International Journal of Machine Learning and Computing, 2011, vol. 1, no 2, pp. 138-147.

17. Kurbatsky V., Sidorov D., Spiryaev V., Tomin N. Forecasting Nonstationary Time Series on the Basis of Hilbert-Huang Transform andMachine Learning. Automation and Remote Control, 2014, vol. 75, no 4, pp.12-16.

18. Leahy P. et al. Structural optimisation and input selection of an artificial neural network for river level prediction. Journal of Hydrology, 2008, vol. 355, pp. 192-201.

19. Louppe G. et al. Understanding variable importances in forests of randomized trees. Advances in Neural Information Processing Systems, 2013, pp. 431-439.

20. Ludermir T.B. et al. An optimization methodology for neural network weights and architectures. IEEE Trans. on Neural Networks, 2006, vol. 17, pp. 1452-1459.

21. Mercer J. Functions of positive and negative type, and their connection with the theory of integral equations. Phil. Trans. of the Royal Society of London. Series A, Containing Papers of a Mathematical or Physical Character, 1909, vol. 209, pp.415-446.

22. Natenberg E.J. Gagne II D.J., Zack J.W., Manobianco J., Van Knowe G.E., Melino T. Application of a Random Forest Approach to Model Output Statistics for use in Day Ahead Wind Power Forecasts. Proc. the Symposium on the Role of Statistical Methods in Weather and Climate Prediction, USA, Austin, 2013.

23. Negnevitsky M., Voropai N., Kurbatsky V., Tomin N., and D. Panasetsky Development of an Intelligent System for Preventing Large-Scale Emergencies in Power Systems. IEEE/PES General Meeting, Vancouver, BC, Canada, 21-25 July2013, pp. 1-6.

24. Neville P.G. Controversy of Variable Importance in Random Forests. Journal of Unified Statistical Techniques, 2013, vol. 1, no 1, pp. 15-20

25. Prechelt L. Proben1 - A set of neural network benchmark problems and benchmark rules. Computer Science Faculty, University of Karlsruhe, Germany, Tech.l Rep. 21/94, Sept. 1994.

26. Sinha N., Lai L.L., Kumar Ghosh P., Ma Y.Wavelet-GA-ANN Based HybridModel for Accurate Prediction of Short-Term Load Forecast. Proc. the IEEE Inter. Conf. on ISAP, 2007, Toki Messe, Niigata, pp. 1-8.

27. Sidorov D. N. Methods of Analysis of Integral Dynamical Models: Theory and Applications (Russian). ISU Publ., Irkutsk, 2013, 293 p.

28. Torres M. E. et al. A complete ensemble empirical mode decomposition with adaptive noise. IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), IEEE, 2011, pp. 4144-4147.

29. Voropai N.I. Glazunova A.M., Kurbatsky V.G., Sidorov D.N., Spiryaev V.A., Tomin N.V. Operating Conditions Forecasting for Monitoring and Control of Electric Power Systems. Proc. the IEEE ISGT Europe 2010 Conference,Gothenburg, Sweden, 2010.

30. Watters C.S., Leahy P. Comparison of linear, Kalman filter and neural network downscaling of wind speeds from numerical weather prediction. Proc. 2011 10th International Conference on Environment and Electrical Engineering, Rome, Italia, 2009.

31. Wu Z., Huang N. E. Ensemble empirical mode decomposition: a noise-assisted data analysis method. Advances in adaptive data analysis, 2009, vol. 1, no. 1, pp. 1-41.

32. Zhu C., Byrd R.H., Lu P., Nocedal J. Algorithm 778: L-BFGS-B, Fortran subroutines for large scale bound constrained optimization. ACM Transactions on Mathematical Software, 1997, vol. 23, no 4, pp. 550-560.