Robust Bayesian Subspace Identification for Small Data Sets

TL;DR

This paper introduces Bayesian and shrinkage estimators for subspace identification, significantly reducing estimation variance in small data sets and large models, which was not previously explored.

Contribution

It applies Bayesian and shrinkage estimation techniques to subspace identification, demonstrating substantial variance reduction in small data scenarios.

Findings

Estimation risk reduced by up to 40%

Bayesian estimators outperform traditional methods

Shrinkage estimators also improve accuracy

Abstract

Model estimates obtained from traditional subspace identification methods may be subject to significant variance. This elevated variance is aggravated in the cases of large models or of a limited sample size. Common solutions to reduce the effect of variance are regularized estimators, shrinkage estimators and Bayesian estimation. In the current work we investigate the latter two solutions, which have not yet been applied to subspace identification. Our experimental results show that our proposed estimators may reduce the estimation risk up to $40\%$ of that of traditional subspace methods.