[Un algorithme de prévision SSA amélioré, reposant sur des séries filtrées]
La technique d'analyse du spectre singulier (SSA) est une méthode puissante et non paramétrique dans le domaine de l'analyse des séries temporelles. Elle connaît depuis ces dernières années une popularité croissante en raison de son large éventail d'applications. La prévision récurrente est l'une des plus importantes méthodes de prévision en SSA. Dans ce texte, nous améliorons la précision de ces prévisions récurrentes en introduisant un nouvel algorithme. Dans notre approche, les coefficients récurrents sont engendrés à partir d'une série filtrée qui a un bruit moindre, ce qui permet d'obtenir de meilleures prévisions. Nous comparons cette nouvelle méthode avec celle établie, en la testant sur des applications à diverses séries temporelles, réelles ou simulées. Les résultats confirment que la nouvelle méthode produit des prévisions plus précises.
The Singular Spectrum Analysis (SSA) technique is a non-parametric powerful method in the field of time series analysis whose popularity has increased in recent years owing to its widespread applications. Recurrent forecasting is one of the important forecasting methods in SSA. In this paper, the forecasting accuracy of recurrent forecasts is improved via the introduction of a new recurrent forecasting algorithm. In the novel approach, the recurrent coefficients are generated from the filtered series which has less noise and thus enables one to achieve the better forecasts. The performance of the new method has been compared with the established recurrent forecasting method. The comparison involves applications to various real and simulated time series. The obtained results confirm that the new approach can provide more accurate forecasts.
Accepté le :
Publié le :
@article{CRMATH_2017__355_9_1026_0,
author = {Hassani, Hossein and Kalantari, Mahdi and Yarmohammadi, Masoud},
title = {An improved {SSA} forecasting result based on a filtered recurrent forecasting algorithm},
journal = {Comptes Rendus. Math\'ematique},
pages = {1026--1036},
publisher = {Elsevier},
volume = {355},
number = {9},
year = {2017},
doi = {10.1016/j.crma.2017.09.004},
language = {en},
url = {http://www.numdam.org/articles/10.1016/j.crma.2017.09.004/}
}
TY - JOUR AU - Hassani, Hossein AU - Kalantari, Mahdi AU - Yarmohammadi, Masoud TI - An improved SSA forecasting result based on a filtered recurrent forecasting algorithm JO - Comptes Rendus. Mathématique PY - 2017 SP - 1026 EP - 1036 VL - 355 IS - 9 PB - Elsevier UR - http://www.numdam.org/articles/10.1016/j.crma.2017.09.004/ DO - 10.1016/j.crma.2017.09.004 LA - en ID - CRMATH_2017__355_9_1026_0 ER -
%0 Journal Article %A Hassani, Hossein %A Kalantari, Mahdi %A Yarmohammadi, Masoud %T An improved SSA forecasting result based on a filtered recurrent forecasting algorithm %J Comptes Rendus. Mathématique %D 2017 %P 1026-1036 %V 355 %N 9 %I Elsevier %U http://www.numdam.org/articles/10.1016/j.crma.2017.09.004/ %R 10.1016/j.crma.2017.09.004 %G en %F CRMATH_2017__355_9_1026_0
Hassani, Hossein; Kalantari, Mahdi; Yarmohammadi, Masoud. An improved SSA forecasting result based on a filtered recurrent forecasting algorithm. Comptes Rendus. Mathématique, Tome 355 (2017) no. 9, pp. 1026-1036. doi : 10.1016/j.crma.2017.09.004. http://www.numdam.org/articles/10.1016/j.crma.2017.09.004/
[1] Singular spectrum analysis of sleep EEG in insomnia, J. Med. Syst., Volume 35 (2011) no. 4, pp. 457-461
[2] Quantifying the correlation between the MEI and LOD variations by decomposing LOD with singular spectrum analysis, Earth on the Edge: Science for a Sustainable Planet, International Association of Geodesy Symposia, vol. 139, 2014, pp. 473-477
[3] Extracting qualitative dynamics from experimental data, Phys. D, Nonlinear Phenom., Volume 20 (1986), pp. 217-236
[4] On the qualitative analysis of experimental dynamical systems (Sarkar, S., ed.), Nonlinear Phenomena and Chaos, Adam Hilger, Bristol, 1986, pp. 113-144
[5] Application of singular spectrum analysis to structural monitoring and damage diagnosis of bridges, Struct. Infrastruct. Eng. Maint. Manag. Life-Cycle Des. Perform., Volume 10 (2014) no. 6, pp. 708-727
[6] Singular spectrum analysis for modeling seasonal signals from GPS time series, J. Geodyn., Volume 72 (2013), pp. 25-35
[7] Time Series Analysis: With Applications in R, Springer, 2008
[8] Forecasting time series with complex seasonal patterns using exponential smoothing, J. Amer. Stat. Assoc., Volume 106 (2011) no. 496, pp. 1513-1527
[9] Singular Spectrum Analysis for Time Series, Springer Briefs in Statistics, Springer, 2013
[10] Singular spectrum analysis: methodology and comparison, J. Data Sci., Volume 5 (2007) no. 2, pp. 239-257
[11] From nature to maths: improving forecasting performance in subspace-based methods using genetics colonial theory, Digit. Signal Process., Volume 51 (2016), pp. 101-109
[12] A preliminary investigation into the effect of outlier(s) on singular spectrum analysis, Fluct. Noise Lett., Volume 13 (2014) no. 14
[13] Forecasting accuracy evaluation of tourist arrivals, Ann. Tour. Res., Volume 63 (2017), pp. 112-127
[14] Forecasting U.S. tourist arrivals using optimal singular spectrum analysis, Tour. Manag., Volume 46 (2015), pp. 322-335
[15] Periodicity of carbon element distribution along casting direction in continuous-casting billet by using singular spectrum analysis, Metall. Mater. Trans. B, Volume 45 (2014) no. 5, pp. 1817-1826
[16] Automatic time series forecasting: the forecast package for R, J. Stat. Softw., Volume 27 (2008) no. 3, pp. 1-22
[17] A state space framework for automatic forecasting using exponential smoothing methods, Int. J. Forecast., Volume 18 (2002) no. 3, pp. 439-454
[18] Singular spectrum analysis for enhancing the sensitivity in structural damage detection, J. Sound Vib., Volume 333 (2014) no. 2, pp. 392-417
[19] Roller element bearing fault diagnosis using singular spectrum analysis, Mech. Syst. Signal Process., Volume 35 (2013) no. 1–2, pp. 150-166
[20] Singular Spectrum Analysis of Biomedical Signals, Taylor & Francis/CRC, 2016
[21] Cross country relations in European tourist arrivals, Ann. Tour. Res., Volume 63 (2017), pp. 151-168
[22] On the use of singular spectrum analysis for forecasting U.S. trade before, during and after the 2008 recession, Int. Econ., Volume 141 (2015), pp. 34-49
Cité par Sources :
