Sample Complexity and Minimax Properties of Exponentially Stable Regularized Estimators

• 成果类型: Article
• 署名作者: Pillonetto, Gianluigi; Scampicchio, Anna
• 署名单位: University of Padua; Swiss Federal Institutes of Technology Domain; ETH Zurich
• 刊物名称: IEEE TRANSACTIONS ON AUTOMATIC CONTROL
• ISSN/ISSBN: 0018-9286
• DOI: 10.1109/TAC.2021.3079296
• 发表日期: 2022
• 页码: 2330-2342
• 关键词: Exponential stability; kernel-based regularization; linear system identification; Minimax Estimation; optimality in order; regularization networks; reproducing kernel hilbert spaces (RKHSs)

摘要

Recent studies have shown how regularization may play an important role in linear system identification. An effective approach consists of searching for the impulse response in a high-dimensional space, e.g., a reproducing kernel Hilbert space (RKHS). Complexity is then controlled using a regularizer, e.g., the RKHS norm, able to encode smoothness and stability information. Examples are RKHSs induced by the so-called stable spline or tuned-correlated kernels, which contain a parameter that regulates impulse response exponential decay. In this article, we derive nonasymptotic upper bounds on the l(2) error of these regularized schemes and study their optimality in order (in the minimax sense). Under white noise inputs and Gaussian measurement noises, we obtain conditions which ensure the optimal convergence rate for all the class of stable spline estimators and several generalizations. Theoretical findings are then illustrated via a numerical experiment.