High-Dimensional Vector Autoregressive Time Series Modeling via Tensor Decomposition
From MaRDI portal
Abstract: The classical vector autoregressive model is a fundamental tool for multivariate time series analysis. However, it involves too many parameters when the number of time series and lag order are even moderately large. This paper proposes to rearrange the transition matrices of the model into a tensor form such that the parameter space can be restricted along three directions simultaneously via tensor decomposition. In contrast, the reduced-rank regression method can restrict the parameter space in only one direction. Besides achieving substantial dimension reduction, the proposed model is interpretable from the factor modeling perspective. Moreover, to handle high-dimensional time series, this paper considers imposing sparsity on factor matrices to improve the model interpretability and estimation efficiency, which leads to a sparsity-inducing estimator. For the low-dimensional case, we derive asymptotic properties of the proposed least squares estimator and introduce an alternating least squares algorithm. For the high-dimensional case, we establish non-asymptotic properties of the sparsity-inducing estimator and propose an ADMM algorithm for regularized estimation. Simulation experiments and a real data example demonstrate the advantages of the proposed approach over various existing methods.
Recommendations
- Factor Models for High-Dimensional Tensor Time Series
- Time-Varying Autoregression with Low-Rank Tensors
- Comments on “Factor Models for High-Dimensional Tensor Time Series”
- Comment on “Factor Models for High-Dimensional Tensor Time Series”
- Comment on “Factor Models for High-Dimensional Tensor Time Series” by Rong Chen, Dan Yang, and Cun-Hui Zhang
- Low Rank and Structured Modeling of High-Dimensional Vector Autoregressions
- Estimation of latent factors for high-dimensional time series
- High dimensional forecasting via interpretable vector autoregression
- Factor models for matrix-valued high-dimensional time series
- Tensor decompositions for learning latent variable models
Cites work
- scientific article; zbMATH DE number 845714 (Why is no real title available?)
- scientific article; zbMATH DE number 2199188 (Why is no real title available?)
- A Multilinear Singular Value Decomposition
- A direct estimation of high dimensional stationary vector autoregressions
- A hybrid bootstrap approach to unit root tests
- A splitting method for orthogonality constrained problems
- Analysis of financial time series
- Asymptotic Theory of Overparameterized Structural Models
- Circular law
- Consistently determining the number of factors in multivariate volatility modelling
- Convex regularization for high-dimensional multiresponse tensor regression
- Cross: efficient low-rank tensor completion
- Dimension Reduction and Coefficient Estimation in Multivariate Linear Regression
- Distributed optimization and statistical learning via the alternating direction method of multipliers
- Estimation and forecasting in vector autoregressive moving average models for rich datasets
- Estimation of (near) low-rank matrices with noise and high-dimensional scaling
- Factor models for matrix-valued high-dimensional time series
- Factor models in high-dimensional time series: A time-domain approach
- Finite-time analysis of vector autoregressive models under linear restrictions
- Forecasting in dynamic factor models subject to structural instability
- Forecasting using a large number of predictors: is Bayesian shrinkage a valid alternative to principal components?
- Global convergence of ADMM in nonconvex nonsmooth optimization
- Joint variable and rank selection for parsimonious estimation of high-dimensional matrices
- Large Bayesian VARMAs
- Low Rank and Structured Modeling of High-Dimensional Vector Autoregressions
- Matrix Variate Regressions and Envelope Models
- Measurement Error in Linear Autoregressive Models
- Minimax sparse principal subspace estimation in high dimensions
- Multivariate reduced-rank regression
- Network vector autoregression
- Oracle inequalities for high dimensional vector autoregressions
- Parametric and semiparametric reduced-rank regression with flexible sparsity
- Performance bounds for parameter estimates of high-dimensional linear models with correlated errors
- Reduced rank stochastic regression with a sparse singular value decomposition
- Reduced rank models for multiple time series
- Reduced rank regression via adaptive nuclear norm penalization
- Regularized estimation in sparse high-dimensional time series models
- SOFAR: Large-Scale Association Network Learning
- STORE: sparse tensor response regression and neuroimaging analysis
- Simultaneous analysis of Lasso and Dantzig selector
- Some results on multivariate autoregressive index models
- Sparse reduced-rank regression for simultaneous dimension reduction and variable selection
- Structured Lasso for regression with matrix covariates
- Supervised singular value decomposition and its asymptotic properties
- Tensor Decompositions and Applications
- Tensor Regression with Applications in Neuroimaging Data Analysis
- Testing and Modeling Multivariate Threshold Models
- Testing for high-dimensional network parameters in auto-regressive models
- Testing for unit roots in autoregressive-moving average models of unknown order
Cited in
(42)- Diagnostic analystics in the Bayesian vector autoregressive model
- Smooth and Probabilistic PARAFAC Model with Auxiliary Covariates
- Simultaneous estimation and group identification for network vector autoregressive model with heterogeneous nodes
- Supervised factor modeling for high-dimensional linear time series
- Bipartite network influence analysis of a two-mode network
- High-dimensional low-rank tensor autoregressive time series modeling
- An Interpretable and Efficient Infinite-Order Vector Autoregressive Model for High-Dimensional Time Series
- Mutual influence regression model
- Reduced Rank Spatio-Temporal Models
- Factor Network Autoregressions
- High-dimensional regularized additive matrix autoregressive model
- Partial Envelope and Reduced-Rank Partial Envelope Vector Autoregressive Models
- Rate-optimal robust estimation of high-dimensional vector autoregressive models
- Reduced-Rank Envelope Vector Autoregressive Model
- Randomized tensor decomposition and optimization in the Tucker and tensor train formats
- Sparse estimators for multivariate integer-valued autoregressive models with applications to inference for Hawkes processes
- Authors' reply to the discussion of: ``New tools for network time series with an application to Covid-19 hospitalisations
- Nuo Xu and Fukang Zhu's contribution to the discussion of: ``New tools for network time series with an application to Covid-19 hospitalisations
- G. Tunnicliffe-Wilson's contribution to the discussion of: ``New tools for network time series with an application to Covid-19 hospitalisations
- Maozai Tian and Tan Meng's contribution to the discussion of: ``New tools for network time series with an application to Covid-19 hospitalisations
- Yanbo Tang's contribution to the discussion of: ``New tools for network time series with an application to Covid-19 hospitalisations
- Alessandra Luati's contribution to the discussion of: ``New tools for network time series with an application to Covid-19 hospitalisations
- Hengxu Liu's contribution to the discussion of: ``New tools for network time series with an application to Covid-19 hospitalisations
- Kuldeep Kumar's contribution to the discussion of: ``New tools for network time series with an application to Covid-19 hospitalisations
- R. Killick, J. Wilson, X. Chen, and R. Lund's contribution to the discussion of: ``New tools for network time series with an application to Covid-19 hospitalisations
- Rajendra Bhansali's contribution to the discussion of: ``New tools for network time series with an application to Covid-19 hospitalisations
- Cristian F. Jiménez-Varón and Marina I. Knight's contribution to the discussion of: ``New tools for network time series with an application to Covid-19 hospitalisations
- Andrej Srakar's contribution to the discussion of `new tools for network time series with an application to Covid-19 hospitalizations' by Nason et al..
- David Hand's contribution to the discussion of: ``New tools for network time series with an application to Covid-19 hospitalisations
- Mirko Armillotta and Konstantinos Fokianos' contribution to the discussion of: ``New tools for network time series with an application to Covid-19 hospitalisations
- Seconder of the vote of thanks to Nason et al. and contribution to the discussion of: ``New tools for network time series with an application to Covid-19 hospitalisations
- Proposer of the vote of thanks to Nason et al. and contribution to the discussion of: ``New tools for network time series with an application to Covid-19 hospitalisations
- A single-mode quasi Riemannian gradient descent algorithm for low-multilinear-rank tensor recovery
- Rank and Factor Loadings Estimation in Time Series Tensor Factor Model by Pre-averaging
- High-dimensional banded vector autoregressions subject to structural breaks
- On a matrix-valued autoregressive model
- Scaled envelope models for multivariate time series
- High-dimensional and banded integer-valued autoregressive processes
- Multivariate spatiotemporal models with low rank coefficient matrix
- Dimension reduction in vector autoregressive models for macroeconomic applications
- Vector autoregression and envelope model
- The coefficient estimation of tensor autoregression based on TR decomposition
This page was built for publication: High-Dimensional Vector Autoregressive Time Series Modeling via Tensor Decomposition
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q5881139)
