Learning linear modules in a dynamic network using regularized kernel-based methods
From MaRDI portal
Publication:2665089
Abstract: In order to identify one system (module) in an interconnected dynamic network, one typically has to solve a Multi-Input-Single-Output (MISO) identification problem that requires identification of all modules in the MISO setup. For application of a parametric identification method this would require estimating a large number of parameters, as well as an appropriate model order selection step for a possibly large scale MISO problem, thereby increasing the computational complexity of the identification algorithm to levels that are beyond feasibility. An alternative identification approach is presented employing regularized kernel-based methods. Keeping a parametric model for the module of interest, we model the impulse response of the remaining modules in the MISO structure as zero mean Gaussian processes (GP) with a covariance matrix (kernel) given by the first-order stable spline kernel, accounting for the noise model affecting the output of the target module and also for possible instability of systems in the MISO setup. Using an Empirical Bayes (EB) approach the target module parameters are estimated through an Expectation-Maximization (EM) algorithm with a substantially reduced computational complexity, while avoiding extensive model structure selection. Numerical simulations illustrate the potentials of the introduced method in comparison with the state-of-the-art techniques for local module identification.
The regularized kernel-based approach (so called EBDM, empirical Bayes direct method) for local module identification in dynamic networks is proposed. The main contribution is the method's robustness in case of unstable modules contributing to the network complex behavior. It is to be noted that proposed approach provides better estimates and a reduced variance thanks to the regularization approach of the method.
Recommendations
- An empirical Bayes approach to identification of modules in dynamic networks
- Kernel methods in system identification, machine learning and function estimation: a survey
- Sparse estimation in linear dynamic networks using the stable spline horseshoe prior
- Learning module networks
- Nonparametric identification of Kronecker networks
Cites work
- A Bayesian approach to sparse dynamic network identification
- A Local Direct Method for Module Identification in Dynamic Networks With Correlated Noise
- A new kernel-based approach for linear system identification
- An empirical Bayes approach to identification of modules in dynamic networks
- ARX modeling of unstable linear systems
- Errors-in-variables identification in dynamic networks-consistency results for an instrumental variable approach
- Hierarchical subspace identification of directed acyclic graphs
- scientific article; zbMATH DE number 44406 (Why is no real title available?)
- scientific article; zbMATH DE number 45871 (Why is no real title available?)
- scientific article; zbMATH DE number 3567782 (Why is no real title available?)
- scientific article; zbMATH DE number 3802652 (Why is no real title available?)
- Identifiability of Dynamical Networks With Partial Node Measurements
- Identifiability of linear dynamic networks
- Identification of dynamic models in complex networks with prediction error methods-basic methods for consistent module estimates
- Identification of Dynamic Models in Complex Networks With Prediction Error Methods: Predictor Input Selection
- Identification of systems with unknown inputs using indirect input measurements
- Kernel methods in system identification, machine learning and function estimation: a survey
- Necessary and Sufficient Conditions for Dynamical Structure Reconstruction of LTI Networks
- On the estimation of transfer functions, regularizations and Gaussian processes-revisited
- Prediction error identification of linear dynamic networks with rank-reduced noise
- Prediction error identification of linear systems: a nonparametric Gaussian regression approach
- Robust EM kernel-based methods for linear system identification
- Signal Selection for Estimation and Identification in Networks of Dynamic Systems: A Graphical Model Approach
- Sparse plus low rank network identification: a nonparametric approach
- Subspace Identification of Large-Scale Interconnected Systems
- Topological Identification in Networks of Dynamical Systems
Cited in
(10)- An empirical Bayes approach to identification of modules in dynamic networks
- Nonparametric identification of Kronecker networks
- Sparse estimation in linear dynamic networks using the stable spline horseshoe prior
- A scalable multi-step least squares method for network identification with unknown disturbance topology
- A frequency domain approach for local module identification in dynamic networks
- Adaptive regularised kernel-based identification method for large-scale systems with unknown order
- Optimal excitation and measurement pattern for cascade networks
- Topology identification of sparse network: a stochastic variational Bayesian approach
- Decentralized diffusion-based learning under non-parametric limited prior knowledge
- Deep networks for system identification: a survey
This page was built for publication: Learning linear modules in a dynamic network using regularized kernel-based methods
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2665089)
