SystemIdentificationResult<T>

Models & Types · AiDotNet.Control

A linear state-space model recovered from measured data.

How It Works

The matrices here are exactly what every other class in this module consumes: hand them to LinearQuadraticRegulator or ModelPredictiveController and you have a controller for a system nobody derived equations for. That path — measure, identify, control — is how most industrial controllers are actually built.

Constructors

Constructor Summary
SystemIdentificationResult(Matrix<T> stateMatrix, Matrix<T> inputMatrix, Vector<T> singularValues, int rank, T residual) Creates a system identification result.

Properties

Property Summary
Matrix<T> InputMatrix { get; } Gets the identified input matrix B.
Matrix<T> StateMatrix { get; } Gets the identified state matrix A.
T Residual { get; } Gets the Frobenius norm of the one-step prediction residual on the training data.
Vector<T> SingularValues { get; } Gets the singular values of the stacked data matrix, largest first.
int Rank { get; } Gets the rank actually used, after truncation.