Struct varpro::solvers::levmar::FitResult

pub struct FitResult<Model, const MRHS: bool>
where
    Model: SeparableNonlinearModel,
    Model::ScalarType: RealField + Scalar + Float,
{
    pub problem: LevMarProblem<Model, MRHS>,
    pub minimization_report: MinimizationReport<Model::ScalarType>,
}

A helper type that contains the fitting problem after the minimization, as well as a report and some convenience functions

Fields

problem: LevMarProblem<Model, MRHS>

the final state of the fitting problem after the minimization finished (regardless of whether fitting was successful or not)

minimization_report: MinimizationReport<Model::ScalarType>

the minimization report of the underlying solver. It contains information about the minimization process and should be queried to see whether the minimization was considered successful

Implementations

impl<Model> FitResult<Model, true>

where Model: SeparableNonlinearModel, Model::ScalarType: RealField + Scalar + Float,

pub fn linear_coefficients( &self ) -> Option<MatrixView<'_, Model::ScalarType, Dyn, Dyn>>

Note This implementation is for fitting problems with multiple right hand sides.

Convenience function to get the linear coefficients after the fit has finished. Will return None if there was an error during fitting.

The coefficients vectors for the individual members of the datasets are the colums of the returned matrix. That means one coefficient vector for each right hand side.

pub fn best_fit(&self) -> Option<OMatrix<Model::ScalarType, Dyn, Dyn>>

Note This implementation is for fitting problems with a single right hand side.

Calculate the values of the model at the best fit parameters. Returns None if there was an error during fitting. Since this is for single right hand sides, the output is a column vector.

impl<Model> FitResult<Model, false>

where Model: SeparableNonlinearModel, Model::ScalarType: RealField + Scalar + Float,

pub fn linear_coefficients( &self ) -> Option<VectorView<'_, Model::ScalarType, Dyn>>

Note This implementation is for fitting problems with a single right hand side.

Convenience function to get the linear coefficients after the fit has finished. Will return None if there was an error during fitting. The coefficients are given as a single vector.

pub fn best_fit(&self) -> Option<OVector<Model::ScalarType, Dyn>>

Note This implementation is for fitting problems with multiple right hand sides

Calculate the values of the model at the best fit parameters. Returns None if there was an error during fitting. Since this is for single right hand sides, the output is a matrix containing composed of column vectors for the right hand sides.

impl<Model, const MRHS: bool> FitResult<Model, MRHS>

where Model: SeparableNonlinearModel, Model::ScalarType: RealField + Scalar + Float,

pub fn nonlinear_parameters(&self) -> OVector<Model::ScalarType, Dyn>

convenience function to get the nonlinear parameters of the model after the fitting process has finished.

pub fn was_successful(&self) -> bool

whether the fit was deemeed successful. The fit might still be not be optimal for numerical reasons, but the minimization process terminated successfully.

Trait Implementations

impl<Model, const MRHS: bool> Debug for FitResult<Model, MRHS>

where Model: SeparableNonlinearModel + Debug, Model::ScalarType: RealField + Scalar + Float + Debug,

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.