OpenANN::StoppingCriteria Class Reference

Stopping criteria for optimization algorithms. More...

#include <StoppingCriteria.h>

Public Member Functions
	StoppingCriteria ()
	Create default stop criteria. More...

Public Attributes
int	maximalFunctionEvaluations
	Maximal number of objective function evaluations. More...
int	maximalIterations
	Maximal number of optimization algorithm iterations. More...
int	maximalRestarts
	Maximal number of optimization algorithm restarts. More...
double	minimalValue
	Minimal objective function value. More...
double	minimalValueDifferences
	Minimal objective function difference between iterations. More...
double	minimalSearchSpaceStep
	Minimal step size in the search step (e. More...

Static Public Attributes
static StoppingCriteria	defaultValue

Detailed Description

Stopping criteria for optimization algorithms.

Note that not every algorithm might check all criteria.

Constructor & Destructor Documentation

OpenANN::StoppingCriteria::StoppingCriteria

(

)

Create default stop criteria.

Member Data Documentation

StoppingCriteria OpenANN::StoppingCriteria::defaultValue

static

int OpenANN::StoppingCriteria::maximalFunctionEvaluations

Maximal number of objective function evaluations.

In backpropagation based optimization the objective function is the error function.

int OpenANN::StoppingCriteria::maximalIterations

Maximal number of optimization algorithm iterations.

int OpenANN::StoppingCriteria::maximalRestarts

Maximal number of optimization algorithm restarts.

double OpenANN::StoppingCriteria::minimalSearchSpaceStep

Minimal step size in the search step (e.

g. gradient norm).

double OpenANN::StoppingCriteria::minimalValue

Minimal objective function value.

double OpenANN::StoppingCriteria::minimalValueDifferences

Minimal objective function difference between iterations.

---

The documentation for this class was generated from the following file:

• OpenANN/optimization/StoppingCriteria.h