×

Maximum a-posteriori estimation of autoregressive processes based on finite mixtures of scale-mixtures of skew-normal distributions. (English) Zbl 07191989

Summary: This article investigates maximum a-posteriori (MAP) estimation of autoregressive model parameters when the innovations (errors) follow a finite mixture of distributions that, in turn, are scale-mixtures of skew-normal distributions (SMSN), an attractive and extremely flexible family of probabilistic distributions. The proposed model allows to fit different types of data which can be associated with different noise levels, and provides a robust modelling with great flexibility to accommodate skewness, heavy tails, multimodality and stationarity simultaneously. Also, the existence of convenient hierarchical representations of the SMSN random variables allows us to develop an EM-type algorithm to perform the MAP estimates. A comprehensive simulation study is then conducted to illustrate the superior performance of the proposed method. The new methodology is also applied to annual barley yields data.

MSC:

62-XX Statistics

Software:

sn; mixsmsn
PDFBibTeX XMLCite
Full Text: DOI

References:

[1] Böhning D. Computer-assisted analysis of mixtures and applications. Meta-analysis, disease mapping and others. Boca Raton, FL: Chapman & Hall/CRC; 2000. [Google Scholar] · Zbl 0951.62088
[2] McLachlan GJ, Peel D. Finite mixture models. Chichester: Wiley; 2000. [Crossref], [Google Scholar] · Zbl 0963.62061
[3] Frühwirth-Schnatter S. Finite mixture and Markov switching models. Springer series in statistics. New York: Springer; 2006. [Google Scholar] · Zbl 1108.62002
[4] Mengersen K, Robert C, Titterington D. Mixtures: estimation and applications. Chichester: Wiley; 2011. [Crossref], [Google Scholar] · Zbl 1218.62003
[5] Punzo A, McNicholas PD.Parsimonious mixtures of multivariate contaminated normal distributions. Biometr J. To appear; 2016. [Crossref], [PubMed], [Web of Science ®], [Google Scholar] · Zbl 1353.62124
[6] Andrews DR, Mallows CL.Scale mixture of normal distributions. J R Statist Soc Ser B. 1974;36:99-102. [Google Scholar] · Zbl 0282.62017
[7] Lin TI, Lee JC, Yen SY.Finite mixture modeling using the skew-normal distribution. Statist Sin. 2007;17(b):909-927. [Google Scholar] · Zbl 1133.62012
[8] Branco, MD, Dey, DK.A general class of multivariate skew-elliptical distributions. J Multivariate Anal. 2001;79:99-113. doi:10.1006/jmva.2000.1960[Crossref], [Web of Science ®], [Google Scholar] · Zbl 0992.62047
[9] Lin TI.Maximum likelihood estimation for multivariate skew normal mixture models. J Multivariate Anal. 2009;100(2):257-265. doi:10.1016/j.jmva.2008.04.010[Crossref], [Web of Science ®], [Google Scholar] · Zbl 1152.62034
[10] Lin TI.Robust mixture modeling using multivariate skew t distributions. Stat Comput. 2010;20(3):343-356. doi:10.1007/s11222-009-9128-9[Crossref], [Web of Science ®], [Google Scholar]
[11] Lee SX, McLachlan GJ.Model-based clustering and classification with non-normal mixture distributions. Statist Methods Appl. 2013;22(4):427-454. doi:10.1007/s10260-013-0237-4[Crossref], [Web of Science ®], [Google Scholar] · Zbl 1332.62209
[12] Lee SX, McLachlan GJ.On mixtures of skew normal and skew t distributions. Adv Data Anal Classif. 2013;7(3):241-266. doi:10.1007/s11634-013-0132-8[Crossref], [Web of Science ®], [Google Scholar] · Zbl 1273.62115
[13] Azzalini A. with the collaboration of Capitanio A. The skew-normal and related families. Cambridge: Cambridge University Press; 2014. IMS Monographs series. [Google Scholar] · Zbl 1338.62007
[14] Genton MG. Skew-elliptical distributions and their applications: a journey beyond normality. Boca Raton, FL: Chapman & Hall/CRC; 2004. Edited Volume. doi:10.1201/9780203492000[Crossref], [Google Scholar] · Zbl 1069.62045
[15] Arellano-Valle RB, Genton MG.On fundamental skew distributions. J Multivariate Anal. 2005;96:93-116. doi:10.1016/j.jmva.2004.10.002[Crossref], [Web of Science ®], [Google Scholar] · Zbl 1073.62049
[16] Arellano-Valle RB, Azzalini A.On unification of skew normal families. Scand J Stat. 2006;33:561-574. doi:10.1111/j.1467-9469.2006.00503.x[Crossref], [Web of Science ®], [Google Scholar] · Zbl 1117.62051
[17] Arellano-Valle RB, Branco MD, Genton MG.A unified view on skewed distributions arising from selections. Canad J Stat. 2006;34:561-574. doi:10.1111/j.1467-9469.2006.00503.x[Crossref], [Web of Science ®], [Google Scholar] · Zbl 1121.60009
[18] Basso RM, Lachos VH, Cabral CRB, Ghosh P.Robust mixture modeling based on the scale mixtures of skew-normal distributions. Comput Stat Data Anal. 2010;54:2926-2941. doi:10.1016/j.csda.2009.09.031[Crossref], [Web of Science ®], [Google Scholar] · Zbl 1284.62193
[19] Pyne S, Hu X, Wang K, Rossin E, Lin TI, Maier LM, Baecher-Allan C, McLachlan GJ, Tamayo P, Hafler DA, De Jager PL, Mesirow JP.Automated high-dimensional flow cytometric data analysis. Proc Natl Acad Sci USA. 2009;106:8519-8524. doi:10.1073/pnas.0903028106[Crossref], [PubMed], [Web of Science ®], [Google Scholar]
[20] Frühwirth-Schnatter S, Pyne S.Bayesian inference for finite mixtures of univariate and multivariate skew-normal and skew-t distributions. Biostatistics. 2010;11(2):317-336. doi:10.1093/biostatistics/kxp062[Crossref], [PubMed], [Web of Science ®], [Google Scholar] · Zbl 1437.62465
[21] Franczak BC, Browne RP, McNicholas PD.Mixtures of shifted asymmetric Laplace distributions. IEEE Trans Pattern Anal. 2014;36(6):1149-1157. doi:10.1109/TPAMI.2013.216[Crossref], [PubMed], [Web of Science ®], [Google Scholar]
[22] Wang K., Ng SK, McLachlan GJ. Multivariate skew t mixture models: Applications to fluorescence-activated cell sorting data. In: Shi H, Zhang Y, Bottema MJ, et al., editors. Proceedings of conference of digital image computing: techniques and applications. Los Alamitos, CA: IEEE; 2009. p. 526-531 [Google Scholar]
[23] Lee SX, McLachlan GJ.Finite mixtures of multivariate skew t distributions: some recent and new results. Stat Comput. 2014;24:181-202. doi:10.1007/s11222-012-9362-4[Crossref], [Web of Science ®], [Google Scholar] · Zbl 1325.62107
[24] Vrbik I, McNicholas PD.Analytic calculations for the EM algorithm for multivariate skew-t mixture models. Stat Probabil Lett. 2012;82(6):1169-1174. doi:10.1016/j.spl.2012.02.020[Crossref], [Web of Science ®], [Google Scholar] · Zbl 1244.65012
[25] Karlis D, Santourian A.Model-based clustering with non-elliptically con-toured distributions. Stat Comput. 2009;19(1):73-83. doi:10.1007/s11222-008-9072-0[Crossref], [Web of Science ®], [Google Scholar]
[26] Morris K, McNicholas P, Punzo A, Browne RP. Robust asymmetric clustering. arXiv.org e-print 1402.6744. 2014; available at: http://arxiv.org/abs/1402.6744[Google Scholar]
[27] Browne RP, McNicholas PD.A mixture of generalized hyperbolic distributions. Canad J Stat. 2015;43(2):176-198. doi:10.1002/cjs.11246[Crossref], [Web of Science ®], [Google Scholar] · Zbl 1320.62144
[28] Li WK, Mcleod AI.ARMA modeling with non-Gaussian innovations. J Time Ser Anal. 1988;9:155-168. doi:10.1111/j.1467-9892.1988.tb00461.x[Crossref], [Google Scholar] · Zbl 0637.62079
[29] Andel J.On AR(1) processes with exponential white noise. Commun Stat - Theory Method. 1988;17:1481-1495. doi:10.1080/03610928808829693[Taylor & Francis Online], [Web of Science ®], [Google Scholar] · Zbl 0639.62082
[30] Janacek GJ, Swift AL.A class of models for non-normal time series. J Time Ser Anal. 1990;11:19-31. doi:10.1111/j.1467-9892.1990.tb00039.x[Crossref], [Google Scholar]
[31] Zellner A.Bayesian and non-Bayesian analysis of the regression model with multivariate student-t error terms. J Am Stat Assoc. 1976;71:400-405. [Taylor & Francis Online], [Web of Science ®], [Google Scholar] · Zbl 0348.62026
[32] Lange KL, Little RJA, Taylor JMG.Robust statistical modeling using the t-distribution. J Am Stat Assoc. 1989;84:881-896. [Taylor & Francis Online], [Web of Science ®], [Google Scholar]
[33] Tiku ML, Wong WK, Vaughan DC, Bian G.Time series models in non-normal situations; Symmetric innovations. J Time Ser Anal. 2000;21:571-596. doi:10.1111/1467-9892.00199[Crossref], [Web of Science ®], [Google Scholar] · Zbl 0972.62084
[34] Penny WD, Roberts SJ.Variational Bayes for generalized autoregressive models. IEEE Transm Signal Process. 2002;50(9):2245-2257. doi:10.1109/TSP.2002.801921[Crossref], [Web of Science ®], [Google Scholar] · Zbl 1369.94269
[35] Christmas J, Everson R.Robust autoregression: Student-t innovations using variational Bayes. IEEE Transm Signal Process. 2011;59:48-57. doi:10.1109/TSP.2010.2080271[Crossref], [Web of Science ®], [Google Scholar] · Zbl 1392.94151
[36] Azzalini A.A class of distributions which includes the normal ones. Scandi J Stat. 1985;12:171-178. [Web of Science ®], [Google Scholar] · Zbl 0581.62014
[37] Pourahmadi M.Skew-normal ARMA models with nonlinear heteroscedastic predictors. Commun Stat Theory Methods. 2007;36:1803-1819. doi:10.1080/03610920601126274[Taylor & Francis Online], [Web of Science ®], [Google Scholar] · Zbl 1315.62073
[38] Azzalini A, Dalla Valle A.The multivariate skew-normal distribution. Biometrika. 1996;83:715-726. doi:10.1093/biomet/83.4.715[Crossref], [Web of Science ®], [Google Scholar] · Zbl 0885.62062
[39] Azzalini A, Capitanio A.Statistical applications of the multivariate skew-normal distribution. J R Stat Soc Ser B. 1999;61:579-602. doi:10.1111/1467-9868.00194[Crossref], [Google Scholar] · Zbl 0924.62050
[40] Minozzo M, Ferracuti L.On the existence of some skew-normal stationary processes. Chilean J Stat. 2012;3(2):157-170. [Google Scholar] · Zbl 1449.62205
[41] Bondon P.Estimation of autoregressive models with epsilon-skew-normal innovations. J Multivariate Anal. 2009;100:1761-1776. doi:10.1016/j.jmva.2009.02.006[Crossref], [Web of Science ®], [Google Scholar] · Zbl 1163.62343
[42] Mudholkar GS, Hutson AD.The epsilon-skew-normal distribution for analyzing near-normal data. J Stat Plan Inference. 2000;83:291-309. doi:10.1016/S0378-3758(99)00096-8[Crossref], [Web of Science ®], [Google Scholar] · Zbl 0943.62012
[43] Hutson A.Utilizing the flexibility of the epsilon-skew-normal distribution for common regression. J Appl Stat. 2004;31(6):673-683. doi:10.1080/1478881042000214659[Taylor & Francis Online], [Web of Science ®], [Google Scholar]
[44] Maleki M, Nematollahi AR.Bayesian approach to epsilon-skew-normal family. Commun Stat Theor Methods. 2016; Just accepted. [Taylor & Francis Online], [Web of Science ®], [Google Scholar] · Zbl 1376.62016
[45] Contreras-Reyes JE.Analyzing fish condition factor index through skew-gaussian information theory quantifiers. Fluctuat Noise Lett. 2016;15(2):1650013. doi:10.1142/S0219477516500139[Crossref], [Web of Science ®], [Google Scholar]
[46] Arellano-Valle RB, Bolfarine H, Lachos VH.Skew-normal linear mixed models. J Data Sci. 2005;3:415-438. [Google Scholar]
[47] Arellano-Valle RB, Ozan S, Bolfarine H, Lachos VH.Skew normal measurement error models. J Multivariate Anal. 2005;96:265-281. doi:10.1016/j.jmva.2004.11.002[Crossref], [Web of Science ®], [Google Scholar] · Zbl 1077.62043
[48] Arellano-Valle RB, Castro LM, Genton MG, Gómez HW.Bayesian inference for shape mixtures of skewed distributions, with application to regression analysis. Bayesian Anal. 2008;3(3):513-539. [Web of Science ®], [Google Scholar] · Zbl 1330.62242
[49] Lachos VH, Ghosh P, Arellano-Valle RB.Likelihood based inference for skew-normal/independent linear mixed models. Statist Sin. 2010a;20:303-322. [Web of Science ®], [Google Scholar] · Zbl 1186.62071
[50] Zeller CB, Cabral CRB, Lachos VH.Robust mixture regression modeling based on scale mixtures of skew-normal distributions. Comput Stat Data Anal. 2010;54:2926-2941. doi:10.1016/j.csda.2009.11.008[Crossref], [Web of Science ®], [Google Scholar] · Zbl 1284.62193
[51] Zeller CB, Lachos VH, Vilca-Labra FE.Local influence analysis for regression models with scale mixtures of skew-normal distributions. J Appl Stat. 2011;38:348-363. doi:10.1080/02664760903406504[Taylor & Francis Online], [Web of Science ®], [Google Scholar] · Zbl 1511.62164
[52] Cancho VG, Dey, DK, Lachos VH, Andrade MG.Bayesian nonlinear regression models with scale mixtures of skew-normal distributions: Estimation and case influence diagnostics. Comput Stat Data Anal. 2011;55:588-602. doi:10.1016/j.csda.2010.05.032[Crossref], [Web of Science ®], [Google Scholar] · Zbl 1247.62083
[53] Garay AM, Lachos VH, Abanto-Valle CA.Nonlinear regression models based on scale mixtures of skew-normal distributions. J Korean Statist Soc. 2011;40:115-124. doi:10.1016/j.jkss.2010.08.003[Crossref], [Web of Science ®], [Google Scholar] · Zbl 1296.62141
[54] Cabral CRB, Lachos VH, Prates MO.Multivariate mixture modeling using skew-normal independent distributions. Comput Stat Data Anal. 2012;56:126-142. doi:10.1016/j.csda.2011.06.026[Crossref], [Web of Science ®], [Google Scholar] · Zbl 1239.62058
[55] Lachos VH, Labra FV, Bolfarine H, Ghosh P.Multivariate measurement error models based on scale mixtures of the skew-normal distribution. Statistics. 2010b;44:541-556. doi:10.1080/02331880903236926[Taylor & Francis Online], [Web of Science ®], [Google Scholar] · Zbl 1291.62120
[56] Labra FV, Garay AM, Lachos VH, Ortega EMM.Estimation and diagnostics for heteroscedastic nonlinear regression models based on scale mixtures of skew-normal distributions. J Statist Plan Inference. 2012;142(7):2149-2165. doi:10.1016/j.jspi.2012.02.018[Crossref], [Web of Science ®], [Google Scholar] · Zbl 1408.62123
[57] Contreras-Reyes JE, Arellano-Valle RB.Growth estimates of cardinalfish (Epigonus crassicaudus) based on scale mixtures of skew-normal distributions. Fish Res. 2013;147:137-144. doi:10.1016/j.fishres.2013.05.002[Crossref], [Web of Science ®], [Google Scholar]
[58] Contreras-Reyes JE, Palma W.Statistical analysis of autoregressive fractionally integrated moving average models in R. Comput Stat. 2013;28(5):2309-2331. doi:10.1007/s00180-013-0408-7[Crossref], [Web of Science ®], [Google Scholar] · Zbl 1306.65051
[59] Contreras-Reyes JE, Arellano-Valle RB, Canales TM.Comparing growth curves with asymmetric heavy-tailed errors: application to the southern blue whiting (Micromesistius australis). Fish Res. 2014;159:88-94. doi:10.1016/j.fishres.2014.05.006[Crossref], [Web of Science ®], [Google Scholar]
[60] Garay AM, Lachos VH, Labra FV, Ortega EMM.Statistical diagnostics for nonlinear regression models based on scale mixtures of skew-normal distributions. J Stat Comput Simul. 2014;84:1761-1778. doi:10.1080/00949655.2013.766188[Taylor & Francis Online], [Web of Science ®], [Google Scholar] · Zbl 1453.62562
[61] Gauvain JL, Lee CH.Maximun a posteriori estimation for multivariate Gaussian mixture observations of Markov chain. IEEE Tran Trans Speech Audio Process 1994;2(2):291-298. doi:10.1109/89.279278[Crossref], [Google Scholar]
[62] Tolpin D, Wood F. Maximun a posteriori estimation by search in probabilistic programs. In Proceedings of the Eighth Annual Symposium on Combinatorial Search; 2015:201-205. [Google Scholar]
[63] Henze N.A probabilistic representation of the skew-normal distribution. Scandi J Stat. 1986;13:271-275. [Web of Science ®], [Google Scholar] · Zbl 0648.62016
[64] Azzalini A, Genton M.Robust likelihood methods based on the skew-t and related distributions. Int Stat Rev. 2008;76:1490-1507. doi:10.1111/j.1751-5823.2007.00016.x[Crossref], [Web of Science ®], [Google Scholar] · Zbl 1206.62102
[65] Arellano-Valle RB, Contreras-Reyes JE, Genton MG.Shannon entropy and mutual information for multivariate skew-elliptical distributions. Scand J Stat. 2013;40:42-46. doi:10.1111/j.1467-9469.2011.00774.x[Crossref], [Web of Science ®], [Google Scholar] · Zbl 1259.62002
[66] Ó’Ruanaidh JJK, Fitzgerald W. Numerical Bayesian methods applied to signal processing. New York: Springer; 1996. [Crossref], [Google Scholar] · Zbl 0871.62025
[67] Liu CH, Rubin DB.The ECME algorithm: a simple extension of EM and ECM with faster monotone convergence. Biometrika. 1994;81:633-648. doi:10.1093/biomet/81.4.633[Crossref], [Web of Science ®], [Google Scholar] · Zbl 0812.62028
[68] Liu CH, Rubin DB.ML estimation of the t distribution using EM and its extensions, ECM and ECME. Statist Sinica. 1995;5:19-40. [Web of Science ®], [Google Scholar] · Zbl 0824.62047
[69] Wei GCG, Tanner MA.A Monte Carlo implementation of the EM algorithm and the poor man’s data augmentation algorithms. J Am Stat Assoc. 1990;85:699-704. doi:10.1080/01621459.1990.10474930[Taylor & Francis Online], [Web of Science ®], [Google Scholar]
[70] McLachlan GJ, Krishnan T. The EM algorithm and extensions. NJ: Wiley; 2008. [Crossref], [Google Scholar] · Zbl 1165.62019
[71] Prates M, Lachos V, Cabral C. Mixsmsn: Fitting finite mixture of scale mixture of skew-normal distributions. R package version 0.3-2; 2011. Available at http://CRAN.R-project.org/package=mixsmsn[Google Scholar]
[72] Akaike H.A new look at the statistical model identification. IEEE Trans Automat Control. 1974;19:716-723. doi:10.1109/TAC.1974.1100705[Crossref], [Web of Science ®], [Google Scholar] · Zbl 0314.62039
[73] Schwarz G.Estimating the dimension of a model. Ann Stat. 1978;6:461-464. doi:10.1214/aos/1176344136[Crossref], [Web of Science ®], [Google Scholar] · Zbl 0379.62005
[74] Brockwell PJ, Davis RA. Time series: theory and methods. 2nd ed. New York: Springer-Verlag; 1991. [Crossref], [Google Scholar] · Zbl 0709.62080
This reference list is based on information provided by the publisher or from digital mathematics libraries. Its items are heuristically matched to zbMATH identifiers and may contain data conversion errors. In some cases that data have been complemented/enhanced by data from zbMATH Open. This attempts to reflect the references listed in the original paper as accurately as possible without claiming completeness or a perfect matching.