shin/NeuralNetworkLib
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

cascade correlation: refactoring

Browse Source
This commit is contained in:
Shin
2016-05-07 21:17:57 +02:00
parent 36ce3f6463
commit eaafc27211
2 changed files with 61 additions and 58 deletions
Download Patch File Download Diff File Expand all files Collapse all files

94
src/NeuralNetwork/ConstructiveAlgorithms/CascadeCorrelation.cpp
View File

@@ -8,14 +8,13 @@ float CascadeCorrelation::trainOutputs(Cascade::Network &network, const std::vec
FeedForward::Network p(network.getNeuronSize() - outputs-1);
p.appendLayer(outputs);
Learning::QuickPropagation learner(p);
learner.setLearningCoefficient(0.9);
for(std::size_t neuron = 0; neuron < outputs; neuron++) {
p[1][neuron+1].setWeights(network.getNeuron(network.getNeuronSize() - outputs + neuron)->getWeights());
p[1][neuron+1].setActivationFunction(network.getNeuron(network.getNeuronSize() - outputs + neuron)->getActivationFunction());
}
std::cout << p.stringify() << "\n";
//std::cout << p.stringify() << "\n";
std::vector<TrainingPattern> patternsForOutput;
for(auto &pattern:patterns) {
@@ -40,12 +39,12 @@ float CascadeCorrelation::trainOutputs(Cascade::Network &network, const std::vec
}
}
if(lastError - error < _minimalErrorStep) {
if(fabs(lastError - error) < _minimalErrorStep) {
iterWithoutImporvement++;
}else {
iterWithoutImporvement=0;
}
} while (iteration++ < 500 && iterWithoutImporvement < 300);
} while (iteration++ < 1000 && iterWithoutImporvement < 3);
std::cout << "iter: " << iteration << ", error: " << error << ", " << (lastError-error) << "\n";
for(std::size_t neuron = 0; neuron < outputs; neuron++) {
@@ -54,7 +53,7 @@ float CascadeCorrelation::trainOutputs(Cascade::Network &network, const std::vec
return error;
}
void CascadeCorrelation::trainCandidates(Cascade::Network &network, std::shared_ptr<Neuron> &candidate, const std::vector<TrainingPattern> &patterns) {
std::shared_ptr<NeuralNetwork::Neuron> CascadeCorrelation::trainCandidates(Cascade::Network &network, std::vector<std::shared_ptr<Neuron>> &candidates, const std::vector<TrainingPattern> &patterns) {
std::size_t outputs = patterns[0].second.size();
std::vector<TrainingPattern> patternsForOutput;
@@ -78,56 +77,67 @@ void CascadeCorrelation::trainCandidates(Cascade::Network &network, std::shared_
}
std::for_each(meanErrors.begin(), meanErrors.end(), [&patterns](float &n){ n/=patterns.size(); });
std::size_t iterations=0;
std::size_t iterationsWithoutIprovement=0;
float lastCorrelations=0;
float correlation=std::numeric_limits<float>::max();
float bestCorrelation=0;
float lastCorrelation=0;
std::shared_ptr<Neuron> bestCandidate=nullptr;
do {
lastCorrelations=correlation;
lastCorrelation = bestCorrelation;
bool firstStep=true;
for(auto&candidate : candidates) {
float correlation;
std::vector<float>activations;
std::vector<float>correlations(errors[0].size());
std::vector<float>correlationSigns(errors[0].size());
std::vector<float> activations;
std::vector<float> correlations(errors[0].size());
std::vector<float> correlationSigns(errors[0].size());
for(auto &pattern:patternsForOutput) {
activations.push_back((*candidate)(pattern.first));
}
for(std::size_t err=0;err<meanErrors.size();err++) {
for(std::size_t activ=0;activ<activations.size();activ++) {
correlations[err] += activations[activ] * (errors[activ][err] - meanErrors[err]);
for(auto &pattern:patternsForOutput) {
activations.push_back((*candidate)(pattern.first));
}
correlationSigns[err] = correlations[err] > 0? 1.0 : -1.0;
}
correlation = std::accumulate(correlations.begin(), correlations.end(),0.0);
std::vector<float> derivatives(candidate->getWeights().size());
for (std::size_t input=0;input<candidate->getWeights().size();input++) {
float dcdw = 0.0;
for(std::size_t err=0;err<errors.size();err++) {
float thetaO = 0.0;
for(std::size_t meanError = 0; meanError < meanErrors.size(); meanError++) {
(*candidate)(patternsForOutput[err].first);
float derivative = candidate->getActivationFunction().derivatedOutput(candidate->value(), candidate->output());
thetaO+=correlationSigns[meanError] * (errors[err][meanError] - meanErrors [meanError]) * derivative * candidate->weight(input);
for(std::size_t err = 0; err < meanErrors.size(); err++) {
for(std::size_t activ = 0; activ < activations.size(); activ++) {
correlations[err] += activations[activ] * (errors[activ][err] - meanErrors[err]);
}
dcdw += thetaO;
correlationSigns[err] = correlations[err] > 0 ? 1.0 : -1.0;
}
correlation = std::accumulate(correlations.begin(), correlations.end(), 0.0);
std::vector<float> derivatives(candidate->getWeights().size());
for(std::size_t input = 0; input < candidate->getWeights().size(); input++) {
float dcdw = 0.0;
for(std::size_t err = 0; err < errors.size(); err++) {
float thetaO = 0.0;
for(std::size_t meanError = 0; meanError < meanErrors.size(); meanError++) {
(*candidate)(patternsForOutput[err].first);
float derivative = candidate->getActivationFunction().derivatedOutput(candidate->value(), candidate->output());
thetaO += correlationSigns[meanError] * (errors[err][meanError] - meanErrors[meanError]) * derivative * candidate->weight(input);
}
dcdw += thetaO;
}
derivatives[input] = dcdw;
}
for(std::size_t weightIndex = 0; weightIndex < derivatives.size(); weightIndex++) {
candidate->weight(weightIndex) += derivatives[weightIndex] * 0.1;
}
if(firstStep || correlation > bestCorrelation) {
bestCorrelation = correlation;
bestCandidate=candidate;
firstStep=false;
}
derivatives[input]=dcdw;
}
for(std::size_t weightIndex = 0; weightIndex < derivatives.size(); weightIndex++) {
candidate->weight(weightIndex) += derivatives[weightIndex]*0.1;
}
if(correlation+0.0001 <= lastCorrelations) {
if(bestCorrelation <= lastCorrelation) {
iterationsWithoutIprovement++;
} else {
iterationsWithoutIprovement=0;
}
std::cout << correlation << "\n";
} while (iterations ++ < 200 && iterationsWithoutIprovement <3);
std::cout << "iter: " << iterations << ", correlation: " << bestCorrelation << ", " << lastCorrelation << "\n";
return bestCandidate;
}