struct mrpt::math::CLevenbergMarquardtTempl::TResultInfo

Overview

#include <mrpt/math/CLevenbergMarquardt.h>

struct TResultInfo
{
    // fields

    NUMTYPE final_sqr_err = 0;
    NUMTYPE initial_sqr_err = 0;
    size_t iterations_executed = 0;
    VECTORTYPE last_err_vector;
    matrix_t path;
    matrix_t H;
};

Fields

VECTORTYPE last_err_vector

The last error vector returned by the user-provided functor.

matrix_t path

Each row is the value of x after each iteration (the first one is x0, the last one the optimum), followed by the squared error.

matrix_t H

The Gauss-Newton approximation of the Hessian, \(J^\top J\), at the last accepted point.

If the residuals have covariance \(\sigma^2 I\), the covariance of the optimal parameters can be estimated as \(\sigma^2 H^{-1}\).