Cited By
View all- Urošević VM. Savić A(2024)Temporal clustering for accurate short-term load forecasting using Bayesian multiple linear regressionApplied Intelligence10.1007/s10489-024-05887-z55:1Online publication date: 25-Nov-2024
Bayesian variable selection for matrix autoregressive models
Abstract
A Bayesian method is proposed for variable selection in high-dimensional matrix autoregressive models which reflects and exploits the original matrix structure of data to (a) reduce dimensionality and (b) foster interpretability of multidimensional relationship structures. A compact form of the model is derived which facilitates the estimation procedure and two computational methods for the estimation are proposed: a Markov chain Monte Carlo algorithm and a scalable Bayesian EM algorithm. Being based on the spike-and-slab framework for fast posterior mode identification, the latter enables Bayesian data analysis of matrix-valued time series at large scales. The theoretical properties, comparative performance, and computational efficiency of the proposed model is investigated through simulated examples and an application to a panel of country economic indicators.
References
[1]
Ahelegbey D, Billio M, and Casarin RBayesian graphical models for structural vector autoregressive processesJ. Appl. Econom.201631357-386https://onlinelibrary.wiley.com/doi/abs/10.1002/jae.2443
[2]
Bader B and Kolda T Algorithm 862: Matlab tensor classes for fast algorithm prototyping ACM Trans. Math. Softw. 2006 32 635-653
[3]
Bai J and Ng S Determining the number of factors in approximate factor models Econometrica 2002 70 191-221
[4]
Bańbura M, Giannone D, and Reichlin L Large Bayesian vector auto regressions J. Appl. Econom. 2010 25 71-92
[5]
Barrett BE and Gray JB A computational framework for variable selection in multivariate regression Stat. Comput. 1994 4 203-212
[6]
Billio M, Casarin R, Iacopini M, and Kaufmann S Bayesian dynamic tensor regression J. Bus. Econ. Stat. 2022
[7]
Bucci A, Ippoliti L, and Valentini P Comparing unconstrained parametrization methods for return covariance matrix prediction Stat. Comput. 2022 32 90
[8]
Camehl A Penalized estimation of panel vector autoregressive models: a panel lasso approach Int. J. Forecast. 2022 6 66
[9]
Canova F and Ciccarelli MEstimating multicountry VAR modelsInt. Econ. Rev.200950929-959https://onlinelibrary.wiley.com/doi/abs/10.1111/j.1468-2354.2009.00554.x
[10]
Canova F and Ciccarelli M Panel vector autoregressive models: a survey Adv. Econom. 2013 31 66
[11]
Castillo I and van der Vaart A Needles and Straw in a Haystack: posterior concentration for possibly sparse sequences Ann. Stat. 2012 40 2069-2101
[12]
Chen EY and Fan J Statistical inference for high-dimensional matrix-variate factor models J. Am. Stat. Assoc. 2021 66 1-18
[13]
Chen L and Huang JZ Sparse reduced-rank regression with covariance estimation Stat. Comput. 2016 26 461-470
[14]
Chen RHX and Yang D Autoregressive models for matrix-valued time series J. Econom. 2021 222 539-560
[15]
Chen R, Yang D, and Zhang C-H Factor models for high-dimensional tensor time series J. Am. Stat. Assoc. 2022 117 94-116
[16]
Cichocki A Fundamental tensor operations for large-scale data analysis using tensor network formats Multidimens. Syst. Signal Process. 2018 29 921-960
[17]
Forni M, Hallin M, Lippi M, and Reichlin L The generalized dynamic factor model J. Am. Stat. Assoc. 2005 100 830-840
[18]
Gao Z and Tsay RS A two-way transformed factor model for matrix-variate time series Econom. Stat. 2021 6 66
[19]
Gefang D Bayesian doubly adaptive elastic-net lasso for VAR shrinkage Int. J. Forecast. 2014 30 1-11 https://www.sciencedirect.com/science/article/pii/S0169207013000770
[20]
George EI and McCulloch RE Variable selection via Gibbs sampling J. Am. Stat. Assoc. 1993 88 881-889
[21]
George EI and McCulloch RE Approaches for Bayesian variable selection Stat. Sin. 1997 66 339-373
[22]
George EI, Sun D, and Ni S Bayesian stochastic search for VAR model restrictions J. Econom. 2008 142 553-580
[23]
Geyer, C.J.: Computation for the Introduction to MCMC Chapter of Handbook of Markov chain Monte Carlo (2010)
[24]
Gong L and Flegal JM A practical sequential stopping rule for high-dimensional Markov chain Monte Carlo J. Comput. Graph. Stat. 2016 25 684-700
[25]
Gupta A and Nagar DK Matrix Variate Distributions 1999 London Chapman & Hall/CRC
[26]
Hoff PD Separable covariance arrays via the Tucker product, with applications to multivariate relational data Bayesian Anal. 2011 6 179-196
[27]
Hoff PD Multilinear tensor regression for longitudinal relational data Ann. Appl. Stat. 2015 9 1169-1193
[28]
Jones GL and Qin Q Markov chain Monte Carlo in practice Annu. Rev. Stat. Appl. 2022 9 557-578
[29]
Kock A and Callot L Oracle inequalities for high dimensional vector autoregressions J. Econom. 2015 186 325-344
[30]
Kolda TG and Bader BW Tensor decompositions and applications SIAM Rev. 2009 51 455-500
[31]
Koop G and Korobilis D Model uncertainty in panel vector autoregressive models Eur. Econ. Rev. 2016 81 115-131
[32]
Koop G, Pesaran M, and Potter SM Impulse response analysis in nonlinear multivariate models J. Econom. 1996 74 119-147 https://www.sciencedirect.com/science/article/pii/0304407695017534
[33]
Korobilis D Prior selection for panel vector autoregressions Comput. Stat. Data Anal. 2016 101 110-120 https://www.sciencedirect.com/science/article/pii/S0167947316300275
[34]
Korobilis D High-dimensional macroeconomic forecasting using message passing algorithms J. Bus. Econ. Stat. 2021 39 493-504
[35]
Lam C, Yao Q, and Bathia N Estimation of latent factors for high-dimensional time series Biometrika 2011 98 901-918
[36]
Lanne M and Nyberg HGeneralized forecast error variance decomposition for linear and nonlinear multivariate modelsOxf. Bull. Econ. Stat.201678595-603https://onlinelibrary.wiley.com/doi/abs/10.1111/obes.12125
[37]
Lütkepohl H New Introduction to Multiple Time Series Analysis 2005 Berlin Springer
[38]
Meng X-L and Rubin DB Maximum likelihood estimation via the ECM algorithm: a general framework Biometrika 1993 80 267-278 http://www.jstor.org/stable/2337198
[39]
Nakajima J Time-varying parameter var model with stochastic volatility: an overview of methodology and empirical applications Monet. Econ. Stud. 2011 29 107-142
[40]
Ohlson M, Rauf Ahmad M, and von Rosen D The multilinear normal distribution: introduction and some basic properties J. Multivar. Anal. 2013 113 37-47 https://www.sciencedirect.com/science/article/pii/S0047259X11001047
[41]
Park T and Casella G The Bayesian Lasso J. Am. Stat. Assoc. 2008 103 681-686
[42]
Pesaran, M.H., Schuermann, T., Weiner, S.M.: Modeling regional interdependencies using a global error-correcting macroeconometric model. J. Bus. Econ. Stat. 22, 129–162 (2004).
[43]
Polson, N., Scott, J., Clarke, B., Severinski, C.: Shrink Globally, Act Locally: Sparse Bayesian Regularization and Prediction, Vol. 9780199694587 (Oxford University Press, Oxford, 2012)
[44]
Ročková V and George EI Emvs: the EM approach to Bayesian variable selection J. Am. Stat. Assoc. 2014 109 828-846
[45]
Ročková V and George EI The spike-and-slab lasso J. Am. Stat. Assoc. 2018 113 431-444
[46]
Rockova V and McAlinn K Dynamic variable selection with spike-and-slab process priors Bayesian Anal. 2021 16 233-269
[47]
Rothman AJ, Levina E, and Zhu J Sparse multivariate regression with covariance estimation J. Comput. Graph. Stat. 2010 19 947-962
[48]
Samanta S, Khare K, and Michailidis G A generalized likelihood-based Bayesian approach for scalable joint regression and covariance selection in high dimensions Stat. Comput. 2022 32 47
[49]
Song, S., Bickel, P.: Large vector auto regressions. Papers, arxiv. org (2011)
[50]
Tucker LR Some mathematical notes on three-mode factor analysis Psychometrika 1966 51 279-311
[51]
Van Loan C The ubiquitous Kronecker product J Comput. Appl. Math. 2000 123 85-100
[52]
Van Loan, C.F., Pitsianis, N.: Approximation with Kronecker Products, pp. 293–314 (Springer, Dordrecht, 1993).
[53]
Vats D, Flegal JM, and Jones GL Multivariate output analysis for Markov chain Monte Carlo Biometrika 2019 106 321-337
[54]
Vats, D., Flegal, J.M. & Jones, G.L.: Monte Carlo Simulation: Are we there yet?, pp. 1–15. Wiley, New York (2021).
[55]
Wang T, Chen M, Zhao H, and Zhu L Estimating a sparse reduction for general regression in high dimensions Stat. Comput. 2018 28 33-46
[56]
Wang D, Liu X, and Chen R Factor models for matrix-valued high-dimensional time series J. Econom. 2019 208 231-248 https://www.sciencedirect.com/science/article/pii/S0304407618301787
Recommendations
Objective Bayesian analysis for autoregressive models with nugget effects
The conditional autoregressive (CAR) and simultaneous autoregressive (SAR) models both have been used extensively for the analysis of spatial structure underlying lattice data in many areas, such as epidemiology, demographics, economics, and geography. ...
Best Subset Selection for Double-Threshold-Variable Autoregressive Moving-Average Models: The Bayesian Approach
AbstractIn this paper, we propose an effective Bayesian subset selection method for the double-threshold-variable autoregressive moving-average (DT-ARMA) models. The usual complexity of estimation is increased mainly by capturing the correlation between ...
Bayesian Estimation of Motion Vector Fields
A stochastic approach to the estimation of 2D motion vector fields from time-varying images is presented. The formulation involves the specification of a deterministic structural model along with stochastic observation and motion field models. Two ...
Comments
Information & Contributors
Information
Published In
© The Author(s) 2024.
Publisher
Kluwer Academic Publishers
United States
Publication History
Published: 11 March 2024
Accepted: 05 February 2024
Received: 10 August 2023
Author Tags
Qualifiers
- Research-article
Funding Sources
Contributors
Other Metrics
Bibliometrics & Citations
Bibliometrics
Article Metrics
- View Citations1Total Citations
- 0Total Downloads
- Downloads (Last 12 months)0
- Downloads (Last 6 weeks)0
Reflects downloads up to 29 Jan 2025
Other Metrics
Citations
Cited By
View all- Urošević VM. Savić A(2024)Temporal clustering for accurate short-term load forecasting using Bayesian multiple linear regressionApplied Intelligence10.1007/s10489-024-05887-z55:1Online publication date: 25-Nov-2024