On Sparse High-Dimensional Graphical Model Learning For Dependent Time Series

@article{Tugnait2022OnSH,
  title={On Sparse High-Dimensional Graphical Model Learning For Dependent Time Series},
  author={Jitendra Tugnait},
  journal={Signal Process.},
  year={2022},
  volume={197},
  pages={108539}
}

Figures from this paper

References

SHOWING 1-10 OF 67 REFERENCES

Graphical Modeling Of High-Dimensional Time Series

  • Jitendra Tugnait
  • Computer Science, Mathematics
    2018 52nd Asilomar Conference on Signals, Systems, and Computers
  • 2018
TLDR
A novel formulation of joint graphical lasso in frequency domain, suitable for dependent time series, generalizing current time-domain approaches to i.i.d. time series is presented.

Consistency of Sparse-Group Lasso Graphical Model Selection for Time Series

  • Jitendra Tugnait
  • Computer Science, Mathematics
    2020 54th Asilomar Conference on Signals, Systems, and Computers
  • 2020
TLDR
Enough conditions are established for consistency of the inverse PSD estimator resulting from the sparse-group graphical lasso-based approach to inferring the conditional independence graph of a high-dimensional stationary multivariate Gaussian time series.

Learning the Conditional Independence Structure of Stationary Time Series: A Multitask Learning Approach

  • A. Jung
  • Computer Science
    IEEE Transactions on Signal Processing
  • 2015
TLDR
The proposed inference scheme works even for sample sizes much smaller than the number of scalar process components if the true underlying CIG is sufficiently sparse and a theoretical performance analysis provides sufficient conditions on the sample size such that the new method is consistent asymptotically.

Compressive nonparametric graphical model selection for time series

TLDR
This work provides analytical conditions for the method to correctly identify the CIG of a high-dimensional discrete-time Gaussian vector random process from finite-length observations and assumes certain spectral smoothness properties.

On Sparse Complex Gaussian Graphical Model Selection

  • Jitendra Tugnait
  • Computer Science, Mathematics
    2019 IEEE 29th International Workshop on Machine Learning for Signal Processing (MLSP)
  • 2019
TLDR
The focus of this paper is on theoretical analysis of a recently proposed graphical lasso approach based on an $\ell_{1} -$penalized log-likelihood objective function to estimate the sparse inverse covariance matrix.

Sparse permutation invariant covariance estimation

TLDR
A method for constructing a sparse estimator for the inverse covariance (concentration) matrix in high-dimensional settings using a penalized normal likelihood approach and forces sparsity by using a lasso-type penalty is proposed.

Bayesian Structure Learning for Stationary Time Series

TLDR
This work takes a Bayesian approach to structure learning, placing priors on (i) the graph structure and (ii) spectral matrices given the graph, and uses a Whittle likelihood approximation and a conjugate prior to define a hyper complex inverse Wishart on the complex-valued and graph-constrained spectralMatrices.

Learning graphical models for stationary time series

TLDR
This paper describes an algorithm for efficient forecasting for stationary Gaussian time series whose spectral densities factorize in a graphical model and shows how to make use of Mercer kernels in this setting, allowing the ideas to be extended to nonlinear models.

An Efficient Approach to Graphical Modeling of Time Series

TLDR
It is shown that reformulation in terms of a multiple hypothesis test reduces computation time by O(p2) and simulations support the assertion that power levels are attained at least as good as those achieved by Matsuda's much slower approach.

On the Sample Complexity of Graphical Model Selection From Non-Stationary Samples

TLDR
A sufficient condition on the required sample size for accurate graphical model selection based on non-stationary data is derived by analyzing a conceptually simple model selection method.
...