Dans cet article nous proposons un modèle dynamique, non-linéaire, pour prévoir en ligne la consommation d’électricité. Nous présentons une revue des méthodes séquentielles de Monte Carlo utilisées pour le calcul des filtres particulaires. Nous discutons les principaux problèmes qui surviennent lors de l’utilisation de ces filtres et nous décrivons les algorithmes associés aux solutions. Nous introduisons une nouvelle étape qui permet la détection et la suppression automatique des données aberrantes qui conduisent souvent à la dégénerescence de la distribution des particules. Nous appliquons ensuite un algorithme de filtrage particulaire afin d’estimer notre modèle de consommation et comparons les prévisions obtenues aux prévisions opérationnelles utilisées au sein d’EDF.
In this paper, we are interested in the online prediction of the electricity load, within the Bayesian framework of dynamic models. We offer a review of sequential Monte Carlo methods, and provide the calculations needed for the derivation of so-called particles filters. We also discuss the practical issues arising from their use, and some of the variants proposed in the literature to deal with them, giving detailed algorithms whenever possible for an easy implementation. We propose an additional step to help make basic particle filters more robust with regard to outlying observations. Finally we use such a particle filter to estimate a state-space model that includes exogenous variables in order to forecast the electricity load for the customers of the French electricity company Électricité de France and discuss the various results obtained.
Mots clés : modèle dynamique bayésien, filtrage particulaire, méthodes de Monte Carlo séquentielles, prévision de la consommation d’électricité
@article{JSFS_2013__154_2_1_0,
author = {Launay, Tristan and Philippe, Anne and Lamarche, Sophie},
title = {On particle filters applied to electricity load forecasting},
journal = {Journal de la soci\'et\'e fran\c{c}aise de statistique},
pages = {1--36},
publisher = {Soci\'et\'e fran\c{c}aise de statistique},
volume = {154},
number = {2},
year = {2013},
zbl = {1316.62184},
language = {en},
url = {http://www.numdam.org/item/JSFS_2013__154_2_1_0/}
}
TY - JOUR AU - Launay, Tristan AU - Philippe, Anne AU - Lamarche, Sophie TI - On particle filters applied to electricity load forecasting JO - Journal de la société française de statistique PY - 2013 SP - 1 EP - 36 VL - 154 IS - 2 PB - Société française de statistique UR - http://www.numdam.org/item/JSFS_2013__154_2_1_0/ LA - en ID - JSFS_2013__154_2_1_0 ER -
%0 Journal Article %A Launay, Tristan %A Philippe, Anne %A Lamarche, Sophie %T On particle filters applied to electricity load forecasting %J Journal de la société française de statistique %D 2013 %P 1-36 %V 154 %N 2 %I Société française de statistique %U http://www.numdam.org/item/JSFS_2013__154_2_1_0/ %G en %F JSFS_2013__154_2_1_0
Launay, Tristan; Philippe, Anne; Lamarche, Sophie. On particle filters applied to electricity load forecasting. Journal de la société française de statistique, Tome 154 (2013) no. 2, pp. 1-36. http://www.numdam.org/item/JSFS_2013__154_2_1_0/
[1] Sequential MCMC for Bayesian Model Selection, IEEE Higher Order Statistics Workshop, Ceasarea (1999), pp. 130-134
[2] Particle Markov chain Monte Carlo methods, Journal of the Royal Statistical Society: Series B, Volume 72 (2010) no. 3, pp. 269-342 | DOI | Zbl
[3] A Non-Linear Regression Model for Mid-Term Load Forecasting and Improvements in Seasonnality, Proceedings of the 15th Power Systems Computation Conference 2005, Liege Belgium (2005)
[4] Improved particle filter for nonlinear problems, Radar, Sonar and Navigation, IEE Proceedings, Volume 146 (1999) no. 1, pp. 2-7 | DOI
[5] Bayesian Filtering : From Kalman Filters to Particles Filters, and Beyond, 2003
[6] SMC: an efficient algorithm for sequential analysis of state-space models, Journal of the Royal Statistical Society: Series B (2012) | arXiv
[7] Online data processing: comparison of Bayesian regularized particle filters, Electron. J. Statist., Volume 3 (2009), pp. 239-258 | DOI | Zbl
[8] Inference in Hidden Markov Models (Springer Series in Statistics), Springer, 2010, 670 pages (ISBN: 1441923195)
[9] Méthodes de Monte Carlo Séquentielles Adaptatives, Université Pierre et Marie Curie (2009) (Ph. D. Thesis)
[10] Asymptotic Theory of Statistics and Probability, Springer Texts in Statistics, Springer, 2008 (ISBN: 0387759700) | Zbl
[11] Comparison of resampling schemes for particle filtering, Image and Signal Processing and Analysis, 2005. ISPA 2005. Proceedings of the 4th International Symposium on, IEEE (2005), pp. 64-69 | DOI
[12] Sequential Monte Carlo Methods in Practice (Statistics for Engineering and Information Science), Springer, 2001 (ISBN: 0387951466) | MR | Zbl
[13] On Sequential Monte Carlo Sampling Methods for Bayesian Filtering, Statistics and Computing, Volume 10 (2000), pp. 197-208 http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.12.5875
[14] A tutorial on particle filtering and smoothing: fifteen years later, 2011, pp. 656-704 http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.157.772 | Zbl
[15] Series Analysis by State Space Methods, Oxford Statistical Science Series., 2001 | Zbl
[16] An Hourly Periodic State Space Model for Modelling French National Electricity Load, International Journal of Forecasting, Volume 24 (2008) no. 4, pp. 566-587
[17] State-space modelling for high frequency data, Vrije Universiteit Amsterdam (2009) (Ph. D. Thesis)
[18] On Sequential Simulation-Based Methods for Bayesian Filtering (1998) no. CUED/F-INFENG/TR. 310 http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.51.7851 (Technical report)
[19] Following a moving target – Monte Carlo inference for dynamic Bayesian models, Journal of the Royal Statistical Society: Series B (2001), pp. 127-146 | DOI | Zbl
[20] Bayesian Inference in Econometric Models Using Monte Carlo Integration, Econometrica, Volume 57 (1989) no. 6, pp. 1317-1339 | DOI | Zbl
[21] Particle filters for positioning, navigation, and tracking, IEEE Transactions on Signal Processing, Volume 50 (2002) no. 2, pp. 425-437 | DOI
[22] Inference from Iterative Simulation Using Multiple Sequences, Statistical Science, Volume 7 (1992) no. 4, pp. 457-472 | DOI | Zbl
[23] Novel approach to nonlinear/non-Gaussian Bayesian state estimation, Radar and Signal Processing, IEE Proceedings F, Volume 140 (1993) no. 2, pp. 107-113 http://ieeexplore.ieee.org/xpls/abs_all.jsp?arnumber=210672
[24] Self-organizing time series model, Sequential Monte Carlo Methods in Practice, Springer-Verlag (2001), pp. 429-444 | Zbl
[25] Quicksort, The Computer Journal, Volume 5 (1962) no. 1, pp. 10-16 | DOI | Zbl
[26] A basic convergence result for particle filtering, IEEE Transactions on Signal Processing, Volume 56 (2008) no. 4, pp. 1337-1348 | Zbl
[27] A General Convergence Result for Particle Filtering, IEEE Transactions on Signal Processing, Volume 59 (2011) no. 7, pp. 3424-3429 | Zbl
[28] Particle methods for maximum likelihood estimation in latent variable models, Statistics and Computing, Volume 18 (2008) no. 1, pp. 47-57 | DOI
[29] Particle Filtering for Positioning and Tracking Applications, Linköping Universitet, March (2005) (Linköping Studies in Science and Technology. Thesis No 924)
[30] Overview of Sequential Monte Carlo methods for parameter estimation on general state space models, 15th IFAC Symposium on System Identification (SYSID), Saint-Malo, France (2009) (invited paper)
[31] Monte Carlo Filter and Smoother for Non-Gaussian Nonlinear State Space Models, Journal of Computational and Graphical Statistics, Volume 5 (1996) no. 1, pp. 1-25 | DOI
[32] A Self-Organizing State-Space Model, Journal of the American Statistical Association, Volume 93 (1998) no. 443, pp. 1203-1215
[33] Sequential Imputations and Bayesian Missing Data Problems, Journal of the American Statistical Association, Volume 89 (1994) no. 425, pp. 278-288 | DOI | Zbl
[34] Blind Deconvolution via Sequential Imputations, Journal of the American Statistical Association, Volume 90 (1995) no. 430 | DOI | Zbl
[35] Sequential Monte Carlo Methods for Dynamic Systems, Journal of the American Statistical Association, Volume 93 (1998), pp. 1032-1044 http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.56.1897 | Zbl
[36] Monte Carlo Strategies in Scientific Computing, Springer, 2008 | DOI | Zbl
[37] Consistency of the posterior distribution and MLE for piecewise linear regression, Electron. J. Statist., Volume 6 (2012), pp. 1307-1357 | DOI | Zbl
[38] Construction of an informative hierarchical prior distribution. Application to electricity load forecasting, Preprint (2012) (arXiv:1109.4533)
[39] The Internet-Based Teleoperation: Motion and Force Predictions Using the Particle Filter Method, ASME Conference Proceedings, Volume 2010 (2010) no. 44458, pp. 765-771 http://link.aip.org/link/abstract/ASMECP/v2010/i44458/p765/s1 | DOI
[40] WinBUGS - A Bayesian modelling framework: Concepts, structure, and extensibility, Statistics and Computing, Volume 10 (2000) no. 4, pp. 325-337 | DOI
[41] Combined parameter and state estimation in simulation-based filtering, Sequential Monte Carlo Methods in Practice (Freitas, De; Gordon, N. J., eds.), Springer-Verlag, New York, 2001 http://ftp.stat.duke.edu/WorkingPapers/99-14.html | Zbl
[42] Bayesian Core : A Practical Approach to Computational Bayesian Statistics, Springer, 2007 | Zbl
[43] Stabilité et approximations particulaires en filtrage non linéaire, Application au pistage, Université de Rennes 1 (2000) (Ph. D. Thesis)
[44] JAGS: A program for analysis of Bayesian graphical models using Gibbs sampling, Proceedings of the 3rd International Workshop on Distributed Statistical Computing (2003)
[45] Filtering via Simulation: Auxiliary Particle Filters, Journal of the American Statistical Association, Volume 94 (1999) no. 446, pp. 590-599 http://www.jstor.org/stable/2670179 | Zbl
[46] Better Proposal Distributions: Object Tracking Using Unscented Particle Filter, 2001 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR 2001), with CD-ROM, 8-14 December 2001, Kauai, HI, USA, IEEE Computer Society (2001), pp. 786-793
[47] Monte Carlo Statistical Methods, Springer, 2004 | DOI
[48] Methodes de Monte Carlo par chaines de Markov, Economica, 1996 | Zbl
[49] Filtrage non linéaire par noyaux de convolution, Application à un procédé de dépollution biologique, Ecole Nationale Supérieure Agronomique de Montpellier (2004) (Ph. D. Thesis)
[50] Density Estimation for Statistics and Data Analysis (Chapman & Hall/CRC Monographs on Statistics & Applied Probability), Chapman and Hall/CRC, 1986 (ISBN: 0412246201) | Zbl
[51] The Unscented Particle Filter, Advances in Neural Information Processing Systems 13 (2001) http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.32.9011
[52] A Self-Organizing State-Space-Model Approach for Parameter Estimation in Hodgkin-Huxley-Type Models of Single Neurons, PLoS Comput Biol, Volume 8 (2012) no. 3 | DOI
[53] Auxiliary particle filtering : recent developments, Bayesian time series models (Barber, David; Cemgil, A. Taylan; Chiappa, Silvia, eds.), Cambridge University Press, Cambridge, 2011 http://wrap.warwick.ac.uk/37957/
[54] The unscented Kalman filter for nonlinear estimation, Adaptive Systems for Signal Processing, Communications, and Control Symposium 2000. AS-SPCC. The IEEE 2000, IEEE (2000), pp. 153-158 | DOI
