Sharper finite-sample OLS parameter error bounds for linear dynamical systems are obtained via a novel second-order martingale decomposition that achieves instance-optimal CLT rates up to constants in Frobenius norm.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
stat.ML 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
CLT-Optimal Parameter Error Bounds for Linear System Identification
Sharper finite-sample OLS parameter error bounds for linear dynamical systems are obtained via a novel second-order martingale decomposition that achieves instance-optimal CLT rates up to constants in Frobenius norm.