Solutions

struct DataDrivenSolution{T} <: DataDrivenDiffEq.AbstractDataDrivenSolution

The solution to a DataDrivenProblem derived via a certain algorithm. The solution is represented via an Basis, which makes it callable.

Fields

• basis: The basis representation of the solution

• retcode: Returncode

• alg: Algorithm

• out: Original output of the solution algorithm

• prob: Problem

• residuals: Residual sum of squares

• dof: Degrees of freedom

• internal_problem: Internal problem

DDReturnCode

Return code for DataDrivenDiffEq solver results.

The values indicate successful convergence, generic failure, and termination due to iteration, wall-time, absolute-tolerance, or relative-tolerance limits.

get_problem(r)

Returns the original DataDrivenProblem.

get_basis(r)

Returns the recovered Basis.

get_algorithm(r)

Returns the algorithm used to derive the solution.

get_results(r)

Returns the original output of the algorithm.

is_converged(r)

Assert the result of the DataDrivenSolution and returns true if successful, false otherwise.

dof(sol)

Returns the degrees of freedom of the DataDrivenSolution.

rss(sol)

Returns the residual sum of squares of the DataDrivenSolution.

loglikelihood(sol)

Returns the log-likelihood of the DataDrivenSolution assuming a normal distributed error.

nobs(sol)

Returns the number of observations of the DataDrivenSolution.

nullloglikelihood(sol)

Return the null log-likelihood of the DataDrivenSolution. This corresponds to a model only fitted with an intercept and a normal distributed error.

r2(sol)

Return the coefficient of determination of the DataDrivenSolution.

Note

Only implements CoxSnell based on StatsAPI.loglikelihood and StatsAPI.nullloglikelihood.

summarystats(sol)

Returns the summarystats for each row of the error for the DataDrivenSolution.