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

25
include/NeuralNetwork/ConstructiveAlgorithms/CascadeCorrelation.h
View File

@@ -13,8 +13,8 @@ namespace NeuralNetwork {
class CascadeCorrelation { class CascadeCorrelation {
typedef std::pair<std::vector<float>, std::vector<float>> TrainingPattern; typedef std::pair<std::vector<float>, std::vector<float>> TrainingPattern;
public: public:
CascadeCorrelation(std::size_t numberOfCandidate = 2, float maxError = 0.7) : CascadeCorrelation(std::size_t numberOfCandidate = 20, float maxError = 0.7) :
_maxError(maxError), _weightRange(1), _numberOfCandidates(numberOfCandidate), _generator(rand()), _distribution() { _maxError(maxError), _weightRange(0.3), _numberOfCandidates(numberOfCandidate), _generator(rand()), _distribution() {
setWeightRange(_weightRange); setWeightRange(_weightRange);
} }
@@ -28,10 +28,10 @@ namespace NeuralNetwork {
int step = 0; int step = 0;
float error = trainOutputs(network, patterns); float error = trainOutputs(network, patterns);
while(step++ < 20 && error > _maxError) { while(step++ < 15 && error > _maxError) {
std::shared_ptr<Neuron> candidate = createCandidate(network.getNeuronSize() - outputs); std::vector<std::shared_ptr<Neuron>> candidates = createCandidates(network.getNeuronSize() - outputs);
trainCandidates(network, candidate, patterns); std::shared_ptr<Neuron> candidate=trainCandidates(network, candidates, patterns);
addBestCandidate(network, candidate); addBestCandidate(network, candidate);
error = trainOutputs(network, patterns); error = trainOutputs(network, patterns);
@@ -60,7 +60,7 @@ namespace NeuralNetwork {
} }
protected: protected:
float _minimalErrorStep = 0.0005; float _minimalErrorStep = 0.00005;
float _maxError; float _maxError;
float _weightRange; float _weightRange;
std::size_t _numberOfCandidates; std::size_t _numberOfCandidates;
@@ -84,15 +84,11 @@ namespace NeuralNetwork {
float trainOutputs(Cascade::Network &network, const std::vector<TrainingPattern> &patterns); float trainOutputs(Cascade::Network &network, const std::vector<TrainingPattern> &patterns);
void trainCandidates(Cascade::Network &network, std::shared_ptr<Neuron> &candidates, const std::vector<TrainingPattern> &patterns); std::shared_ptr<Neuron> trainCandidates(Cascade::Network &network, std::vector<std::shared_ptr<Neuron>> &candidates, const std::vector<TrainingPattern> &patterns);
void addBestCandidate(Cascade::Network &network, const std::shared_ptr<Neuron> &candidate) { void addBestCandidate(Cascade::Network &network, const std::shared_ptr<Neuron> &candidate) {
auto neuron = network.addNeuron(); auto neuron = network.addNeuron();
//auto tmp = candidate->getWeights();
//std::fill(tmp.begin(),tmp.end(),0.2);
//neuron->setWeights(tmp);
neuron->setWeights(candidate->getWeights()); neuron->setWeights(candidate->getWeights());
neuron->setActivationFunction(candidate->getActivationFunction()); neuron->setActivationFunction(candidate->getActivationFunction());
for(auto &n :network.getOutputNeurons()) { for(auto &n :network.getOutputNeurons()) {
@@ -102,12 +98,10 @@ namespace NeuralNetwork {
} }
weights[weights.size()-1] = _distribution(_generator); weights[weights.size()-1] = _distribution(_generator);
n->setWeights(weights); n->setWeights(weights);
//n->weight(n->getWeights().size() - 1) = _distribution(_generator);
//n->weight(n->getWeights().size() - 1) = 0.2;//.distribution(_generator);
} }
} }
std::shared_ptr<Neuron> createCandidate(std::size_t id) { //TODO std::vector<std::shared_ptr<Neuron>> createCandidates(std::size_t id) {
std::vector<std::shared_ptr<Neuron>> candidates; std::vector<std::shared_ptr<Neuron>> candidates;
for(std::size_t i = 0; i < _numberOfCandidates; i++) { for(std::size_t i = 0; i < _numberOfCandidates; i++) {
@@ -117,10 +111,9 @@ namespace NeuralNetwork {
for(std::size_t weightIndex = 0; weightIndex < id; weightIndex++) { for(std::size_t weightIndex = 0; weightIndex < id; weightIndex++) {
candidates.back()->weight(weightIndex) = _distribution(_generator); candidates.back()->weight(weightIndex) = _distribution(_generator);
// candidates.back()->weight(weightIndex) = 0.1;//_distribution(_generator);
} }
} }
return candidates[0]; return candidates;
} }
}; };

40
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); FeedForward::Network p(network.getNeuronSize() - outputs-1);
p.appendLayer(outputs); p.appendLayer(outputs);
Learning::QuickPropagation learner(p); Learning::QuickPropagation learner(p);
learner.setLearningCoefficient(0.9);
for(std::size_t neuron = 0; neuron < outputs; neuron++) { 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].setWeights(network.getNeuron(network.getNeuronSize() - outputs + neuron)->getWeights());
p[1][neuron+1].setActivationFunction(network.getNeuron(network.getNeuronSize() - outputs + neuron)->getActivationFunction()); 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; std::vector<TrainingPattern> patternsForOutput;
for(auto &pattern:patterns) { 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++; iterWithoutImporvement++;
}else { }else {
iterWithoutImporvement=0; iterWithoutImporvement=0;
} }
} while (iteration++ < 500 && iterWithoutImporvement < 300); } while (iteration++ < 1000 && iterWithoutImporvement < 3);
std::cout << "iter: " << iteration << ", error: " << error << ", " << (lastError-error) << "\n"; std::cout << "iter: " << iteration << ", error: " << error << ", " << (lastError-error) << "\n";
for(std::size_t neuron = 0; neuron < outputs; neuron++) { for(std::size_t neuron = 0; neuron < outputs; neuron++) {
@@ -54,7 +53,7 @@ float CascadeCorrelation::trainOutputs(Cascade::Network &network, const std::vec
return error; 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::size_t outputs = patterns[0].second.size();
std::vector<TrainingPattern> patternsForOutput; std::vector<TrainingPattern> patternsForOutput;
@@ -78,13 +77,17 @@ void CascadeCorrelation::trainCandidates(Cascade::Network &network, std::shared_
} }
std::for_each(meanErrors.begin(), meanErrors.end(), [&patterns](float &n){ n/=patterns.size(); }); std::for_each(meanErrors.begin(), meanErrors.end(), [&patterns](float &n){ n/=patterns.size(); });
std::size_t iterations=0; std::size_t iterations=0;
std::size_t iterationsWithoutIprovement=0; std::size_t iterationsWithoutIprovement=0;
float lastCorrelations=0; float bestCorrelation=0;
float correlation=std::numeric_limits<float>::max(); float lastCorrelation=0;
std::shared_ptr<Neuron> bestCandidate=nullptr;
do { do {
lastCorrelations=correlation; lastCorrelation = bestCorrelation;
bool firstStep=true;
for(auto&candidate : candidates) {
float correlation;
std::vector<float> activations; std::vector<float> activations;
std::vector<float> correlations(errors[0].size()); std::vector<float> correlations(errors[0].size());
@@ -122,12 +125,19 @@ void CascadeCorrelation::trainCandidates(Cascade::Network &network, std::shared_
candidate->weight(weightIndex) += derivatives[weightIndex] * 0.1; candidate->weight(weightIndex) += derivatives[weightIndex] * 0.1;
} }
if(correlation+0.0001 <= lastCorrelations) { if(firstStep || correlation > bestCorrelation) {
iterationsWithoutIprovement++; bestCorrelation = correlation;
} else { bestCandidate=candidate;
iterationsWithoutIprovement=0; firstStep=false;
}
} }
std::cout << correlation << "\n"; if(bestCorrelation <= lastCorrelation) {
} while (iterations ++ < 200 && iterationsWithoutIprovement <3); iterationsWithoutIprovement++;
}
} while (iterations ++ < 200 && iterationsWithoutIprovement <3);
std::cout << "iter: " << iterations << ", correlation: " << bestCorrelation << ", " << lastCorrelation << "\n";
return bestCandidate;
} }