Sample Complexity Bounds for Linear System Identification from a Finite Set

Nicolas Chatzikiriakos, Andrea Iannelli
9/17/2024
eess.SYcs.LGstat.ML

No official code found. Best implementation: col-tasas/2024-bounds-finite-set-ID (2 stars, MIT, last commit 2025-08-15).

Abstract

This paper considers a finite sample perspective on the problem of identifying an LTI system from a finite set of possible systems using trajectory data. To this end, we use the maximum likelihood estimator to identify the true system and provide an upper bound for its sample complexity. Crucially, the derived bound does not rely on a potentially restrictive stability assumption. Additionally, we leverage tools from information theory to provide a lower bound to the sample complexity that holds independently of the used estimator. The derived sample complexity bounds are analyzed analytically and numerically.

View on arXivView PDF

Code Implementations(1)

col-tasas/2024-bounds-finite-set-IDCommunity100%
MIT

This repository contains the code from our paper "Sample Complexity Bounds for Linear System Identification from a Finite Set".

20PythonSep 10, 20241 years agoMIT

Similar Papers

A finite-sample bound for identifying partially observed linear switched systems from a single trajectory

83% match2025

High Effort, Low Gain: Fundamental Limits of Active Learning for Linear Dynamical Systems

79% match1 repo2025

Finite Sample Identification of Partially Observed Bilinear Dynamical Systems

79% match2025

A least-square method for non-asymptotic identification in linear switching control

79% match2024

Frequency Response Identification of Low-Order Systems: Finite-Sample Analysis

78% match1 repo22025

A New Approach to Learning Linear Dynamical Systems

78% match2023

Cite this paper

@article{chatzikiriakos2024sample,
  title  = {Sample Complexity Bounds for Linear System Identification from a Finite Set},
  author = {Nicolas Chatzikiriakos and Andrea Iannelli},
  year   = {2024},
  eprint = {2409.11141},
  archivePrefix = {arXiv},
  url    = {http://arxiv.org/abs/2409.11141v2}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/2409.11141.svg)](https://codeofpaper.com/paper/2409.11141)

Discussion