A hybrid forecasting algorithm based on SVR and wavelet decomposition

Timotheos Paraskevopoulos; Peter N Posch; Timotheos Paraskevopoulos; Peter N Posch

doi:10.3934/QFE.2018.3.525

Quantitative Finance and Economics

2018, Volume 2, Issue 3: 525-553. doi: 10.3934/QFE.2018.3.525

Previous Article Next Article

Research article

A hybrid forecasting algorithm based on SVR and wavelet decomposition

TU Dortmund University, Finance, Otto-Hahn-Str. 6, 44227 Dortmund, Germany

Received: 13 November 2017 Accepted: 12 February 2018 Published: 23 July 2018
JEL Codes: C63, C53

We present a forecasting algorithm based on support vector regression emphasizing the practical benefits of wavelets for financial time series. We utilize an e ective de-noising algorithm based on wavelets feasible under the assumption that the data is generated by a systematic pattern plus random noise. The learning algorithm focuses solely on the time frequency components, instead of the full time series, leading to a more general approach. Our findings propose how machine learning can be useful for data science applications in combination with signal processing methods. The timefrequency decomposition enables the learning algorithm to solely focus on periodical components that are beneficial to the forecasting power as we drop features with low explanatory power. The proposed integration of feature selection and parameter optimization in a single optimization step enable the proposed algorithm to be scaled for a variety of applications. Applying the algorithm to real life financial data shows wavelet decompositions based on the Daubechie and Coiflet basis functions to deliver the best results for the classification task.
- regression,
- spectral analysis,
- time series,
- financial markets
Citation: Timotheos Paraskevopoulos, Peter N Posch. A hybrid forecasting algorithm based on SVR and wavelet decomposition[J]. Quantitative Finance and Economics, 2018, 2(3): 525-553. doi: 10.3934/QFE.2018.3.525

Related Papers:

Abstract

We present a forecasting algorithm based on support vector regression emphasizing the practical benefits of wavelets for financial time series. We utilize an e ective de-noising algorithm based on wavelets feasible under the assumption that the data is generated by a systematic pattern plus random noise. The learning algorithm focuses solely on the time frequency components, instead of the full time series, leading to a more general approach. Our findings propose how machine learning can be useful for data science applications in combination with signal processing methods. The timefrequency decomposition enables the learning algorithm to solely focus on periodical components that are beneficial to the forecasting power as we drop features with low explanatory power. The proposed integration of feature selection and parameter optimization in a single optimization step enable the proposed algorithm to be scaled for a variety of applications. Applying the algorithm to real life financial data shows wavelet decompositions based on the Daubechie and Coiflet basis functions to deliver the best results for the classification task.

References

[1]	Ambroise C, McLachlan GJ (2002) Selection bias in gene extraction on the basis of microarray geneexpression data. Proceedings of the National Academy of Sciences of the United States of America 99: 6562–6566. doi: 10.1073/pnas.102102699
[2]	Atsalakis GS, Valavanis KP (2009) Surveying stock market forecasting techniques - part 2: Soft computing methods. Expert Syst Appl 36: 5932–5941. doi: 10.1016/j.eswa.2008.07.006
[3]	Chang PC, Fan CY (2008) A hybrid system integrating a wavelet and tsk fuzzy rules for stock price forecasting. Systems, Man, and Cybernetics, Part C: Applications and Reviews, IEEE Transactions on 38: 802–815.
[4]	Chang SI, Yadama S (2010) Statistical process control for monitoring non-linear profiles using wavelet filtering and b-spline approximation. Int J Prod Res 48: 1049–1068. doi: 10.1080/00207540802454799
[5]	Cherkassky V, Ma Y (2004) Practical selection of svm parameters and noise estimation for svm regression. Neural networks 17: 113–126. doi: 10.1016/S0893-6080(03)00169-2
[6]	Daubechies I (1992) Ten Lectures on Wavelets, Society for Industrial and Applied Mathematics.
[7]	DeLurgio S (1998) Forecasting principles and applications, Boston: Irwin/McGraw-Hill.
[8]	Doroslovacki M (1998) On the least asymmetric wavelets. IEEE T Signal Proces 46: 1125–1130. doi: 10.1109/78.668562
[9]	Genc¸ay R, Selc¸uk F, Whitcher BJ (2001) An introduction to wavelets and other filtering methods in finance and economics. Academic press.
[10]	Huang SC, Wu TK (2008) Combining wavelet-based feature extractions with relevance vector machines for stock index forecasting. Expert Syst 25: 133–149. doi: 10.1111/j.1468-0394.2008.00443.x
[11]	Ince H, Trafalis TB (2008) Short term forecasting with support vector machines and application to stock price prediction. Int J Gen Syst 37: 677–687. doi: 10.1080/03081070601068595
[12]	John GH, Kohavi R, Pfleger K, et al. (1994) Irrelevant features and the subset selection problem. In Machine Learning: Proceedings of the Eleventh International Conference, 121–129.
[13]	Kao LJ, Chiu CC, Lu CJ, et al. (2013) A hybrid approach by integrating wavelet-based feature extraction with MARS and SVR for stock index forecasting. Decis Support Syst 54: 1228–1244. doi: 10.1016/j.dss.2012.11.012
[14]	Kim KJ, Han I (2000) Genetic algorithms approach to feature discretization in artificial neural networks for the prediction of stock price index. Expert syst Appl 19: 125–132. doi: 10.1016/S0957-4174(00)00027-0
[15]	Li T, Li Q, Zhu S, et al. (2002) A survey on wavelet applications in data mining. ACM SIGKDD Explorations Newsletter 4: 49–68. doi: 10.1145/772862.772870
[16]	Lu CJ, Lee TS, Chiu CC (2009) Financial time series forecasting using independent component analysis and support vector regression. Decis Support Syst 47: 115–125. doi: 10.1016/j.dss.2009.02.001
[17]	Mattera D, Haykin S (1999) Support vector machines for dynamic reconstruction of a chaotic system. In Advances in kernel methods, MIT Press, 211–241.
[18]	Percival DB, Walden AT (2006) In Wavelet methods for time series analysis, Cambridge University Press.
[19]	Ramsey JB (2002) Wavelets in economics and finance: past and future. Stud Nonlinear Dyn E 6: 1090.
[20]	Ramsey JB, Lampart C (1998) The decomposition of economic relationships by time scale using wavelets: expenditure and income. Stud Nonlinear Dyn E 3: 49–71.
[21]	Saeys Y, Inza I, Larra˜naga P (2007) A review of feature selection techniques in bioinformatics. Bioinformatics 23: 2507–2517. doi: 10.1093/bioinformatics/btm344
[22]	Sch¨olkopf B, Smola A (2005) Support vector machines and kernel algorithms. In Encyclopedia of Biostatistics 2. ed., 5328–5335.
[23]	Seo Y, Kim S, Kisi O, et al. (2015) Daily water level forecasting using wavelet decomposition and artificial intelligence techniques. J Hydrol 520: 224–243. doi: 10.1016/j.jhydrol.2014.11.050
[24]	Svetnik V, Liaw A, Tong C, et al. (2004) Application of breimans random forest to modeling structureactivity relationships of pharmaceutical molecules. In Multiple Classifier Systems, 334–343.
[25]	Vapnik V (1999) An overview of statistical learning theory. IEEE T Neural Network 10: 988–999. doi: 10.1109/72.788640
[26]	Vapnik V, Golowich SE, Smola A (1997) Support vector method for function approximation, regression estimation, and signal processing, In M. Mozer and M. Jordan and T. Petsche, Advances in Neural Information Processing Systems, MIT Press, 281–287.
[27]	Wang GJ, Xie C, Chen S (2017) Multiscale correlation networks analysis of the us stock market: a wavelet analysis. J Econ Interact Coordination 12: 561–594. doi: 10.1007/s11403-016-0176-x
[28]	Zhang BL, Coggins R, Jabri MA, et al. (2001) Multiresolution forecasting for futures trading using wavelet decompositions. IEEE T Neural Network 12: 765–775. doi: 10.1109/72.935090
[29]	Zheng G, Starck J, Campbell J, et al. (1999) Multiscale transforms for filtering financial data streams. J Comput Intell Financ 7: 18–35.

Reader Comments

Your name:*

Email:*
© 2018 the Author(s), licensee AIMS Press. This is an open access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0)