de.kumpe.hadooptimizer.examples.neurons
Class NeuronalNetEvaluator

java.lang.Object
  de.kumpe.hadooptimizer.examples.neurons.NeuronalNetEvaluator

All Implemented Interfaces:

Evaluator<double[]>, ReportingHalterWrapper.Reporter<double[]>, Serializable

public class NeuronalNetEvaluator
extends Object
implements Evaluator<double[]>, ReportingHalterWrapper.Reporter<double[]>

See Also:

Serialized Form

Nested Class Summary

static class

NeuronalNetEvaluator.Sample

Constructor Summary

NeuronalNetEvaluator(InputValue[] inputValues, Neuron.Input[] inputs, Value output, NeuronalNetEvaluator.Sample[] samples)

Method Summary

double	evaluate(double[] weights) Calculates the evaluation (fitness) of the given individual.
void	print(double[] weights)
void	report(Collection<EvaluationResult<double[]>> evaluationResults)

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Detail

NeuronalNetEvaluator

public NeuronalNetEvaluator(InputValue[] inputValues,
                            Neuron.Input[] inputs,
                            Value output,
                            NeuronalNetEvaluator.Sample[] samples)

Method Detail

print

public void print(double[] weights)

evaluate

public double evaluate(double[] weights)

Description copied from interface: Evaluator

Calculates the evaluation (fitness) of the given individual.

A smaller value means better fitness.

Specified by:

evaluate in interface Evaluator<double[]>

Parameters:

weights - the individual to calculate the evaluation (fitness)

Returns:

the evaluation (fitness) of the given individual

report

public void report(Collection<EvaluationResult<double[]>> evaluationResults)

Specified by:

report in interface ReportingHalterWrapper.Reporter<double[]>