High-Dimensional Matrix-Variate Diffusion Index Models for Time Series Forecasting

This paper proposes a novel diffusion-index model for forecasting when predictors are high-dimensional matrix-valued time series. We apply an $\alpha$-PCA method to extract low-dimensional matrix factors and build a bilinear regression linking future outcomes to these factors, estimated via iterative least squares. To handle weak factor structures, we introduce a supervised screening step to select informative rows and columns. Theoretical properties, including consistency and asymptotic normality, are established. Simulations and real data show that our method significantly improves forecast accuracy, with the screening procedure providing additional gains over standard benchmarks in out-of-sample mean squared forecast error.

• Publication year

  2025

• Venue

  Unknown venue

• Publication date

  2025-08-06

• Fields of study

  Mathematics, Computer Science, Economics

• Identifiers
  arXiv 2508.04259
• External record

  Open on Semantic Scholar

• Source metadata

  Semantic Scholar

• No claims are published for this paper.

• No concepts are published for this paper.

• Supervised Dynamic PCA: Linear Dynamic Forecasting with Many Predictors

  2023cited by this paper
• Modelling matrix time series via a tensor CP-decomposition

  2021cited by this paper
• Threshold Selection in Feature Screening for Error Rate Control

  2021cited by this paper
• Simultaneous Decorrelation of Matrix Time Series

  2021cited by this paper
• Approximate factor models with weaker loadings

  2021cited by this paper
• A Two-Way Transformed Factor Model for Matrix-Variate Time Series

  2020cited by this paper
• Statistical Inference for High-Dimensional Matrix-Variate Factor Models

  2020influential reference
• Segmenting High-dimensional Matrix-valued Time Series via Sequential Transformations

  2020cited by this paper
• Lasso inference for high-dimensional time series

  2020cited by this paper
• Scaled PCA: A New Approach to Dimension Reduction

  2019cited by this paper
• Shrinking Factor Dimension: A Reduced-Rank Approach

  2019cited by this paper
• Empirical Asset Pricing via Machine Learning

  2018cited by this paper
• A Structural‐Factor Approach to Modeling High‐Dimensional Time Series and Space‐Time Data

  2018cited by this paper
• Autoregressive models for matrix-valued time series

  2018influential reference
• Regime-switching factor models for high-dimensional time series

  2017cited by this paper
• Factor models for matrix-valued high-dimensional time series

  2016cited by this paper
• Efficient dimension reduction for high-dimensional matrix-valued data

  2016cited by this paper
• STRUCTURED LASSO FOR REGRESSION WITH MATRIX COVARIATES

  2014cited by this paper
• High dimensional stochastic regression with latent factors, endogeneity and nonlinearity

  2013cited by this paper
• Eigenvalue Ratio Test for the Number of Factors

  2013cited by this paper
• Multivariate Time Series Analysis: With R and Financial Applications

  2013cited by this paper
• Confidence intervals and hypothesis testing for high-dimensional regression

  2013cited by this paper
• Time Series: Theory and Methods

  2013cited by this paper
• Sparse Matrix Graphical Models

  2012cited by this paper
• Regularized matrix regression

  2012cited by this paper
• Feature Screening via Distance Correlation Learning

  2012cited by this paper
• Model selection and estimation in the matrix normal graphical model

  2012cited by this paper
• Factor modeling for high-dimensional time series: inference for the number of factors

  2012influential reference
• Estimation of latent factors for high-dimensional time series

  2011cited by this paper
• Statistics for High-Dimensional Data: Methods, Theory and Applications

  2011cited by this paper
• HIGH DIMENSIONAL COVARIANCE MATRIX ESTIMATION IN APPROXIMATE FACTOR MODELS.

  2011cited by this paper
• Large covariance estimation by thresholding principal orthogonal complements

  2011cited by this paper
• Least squares after model selection in high-dimensional sparse models

  2009cited by this paper
• Forecasting economic time series using targeted predictors

  2008cited by this paper
• New Introduction to Multiple Time Series Analysis

  2007cited by this paper
• Inferential Theory for Factor Models of Large Dimensions

  2003cited by this paper
• Nber Working Paper Series Measuring the Effects of Monetary Policy : a Factor-augmented Vector Autoregressive ( Favar ) Approach

  2003cited by this paper
• Forecasting Using Principal Components From a Large Number of Predictors

  2002cited by this paper
• Determining the Number of Factors in Approximate Factor Models

  2000cited by this paper
• Regression Shrinkage and Selection via the Lasso

  1996cited by this paper
• Policy analysis with econometric models.

  1996cited by this paper
• Multivariate Simultaneous Generalized ARCH

  1995cited by this paper
• email

  1995cited by this paper
• Comparing predictive accuracy I: an asymptotic test

  1991influential reference
• Model Specification in Multivariate Time Series

  1989cited by this paper
• An Introduction to Multivariate Statistical Analysis

  1986cited by this paper
• Generalized autoregressive conditional heteroskedasticity

  1986cited by this paper
• Autoregressive conditional heteroscedasticity with estimates of the variance of United Kingdom inflation

  1982cited by this paper
• Modeling Multiple Time Series with Applications

  1981cited by this paper
• Time series analysis, forecasting and control

  1972cited by this paper
• ASYMPTOTIC THEORY FOR PRINCIPAL COMPONENT ANALYSIS

  1963cited by this paper
• Journal of the American Statistical Association Martingale Difference Correlation and Its Use in High- Dimensional Variable Screening Martingale Difference Correlation and Its Use in High-dimensional Variable Screening

  year unknowncited by this paper

• No citing papers are available for this paper.