- add Rbm::normalize()

- normalize of training data

git-svn-id: http://moon:8086/svn/software/trunk/projects/Rbm@660 b431acfa-c32f-4a4a-93f1-934dc6c82436

source/MainComponent.hpp

@@ -100,7 +100,7 @@ private:
 
 	void loadTraining()
 	{
-		m_stack->loadTraining();
+		m_stack->loadTraining(rbmNormalizeDataToggleButton->getToggleState());
 		patterSlider->setRange(0, m_stack->numTraining()-1, 1);
 	}
 

source/Rbm.cpp

@@ -243,6 +243,18 @@ arma::mat Rbm::toVisibleProbs(const arma::mat &hidden) const
 	return probsLogistic(toVisibleState(hidden));
 }
 
+arma::mat Rbm::normalize(const arma::mat& src)
+{
+	double mean = arma::accu(src)/src.n_elem;
+	arma::mat x = src - mean;
+	arma::mat x2 = x % x;
+	double stddev = sqrt(arma::accu(x2)/x2.n_elem);
+	
+	std::cout << "mean" << "  : " << std::endl << mean << std::endl;
+	std::cout << "stddev" << ": " << std::endl << stddev << std::endl;
+	return x/stddev;
+}
+
 void Rbm::uniform(arma::mat& srcDst, double stdDev, double mu)
 {
 #if 1

source/Rbm.hpp

@@ -125,6 +125,7 @@ public:
 	arma::mat toVisibleState(const arma::mat &hidden) const;
 	arma::mat toHiddenProbs(const arma::mat &visible) const;
 	arma::mat toVisibleProbs(const arma::mat &hidden) const;
+	static arma::mat normalize(const arma::mat &hidden);
 	const arma::mat& w() const;
 	const arma::mat& bv() const;
 	const arma::mat& bh() const;

source/Stack.cpp

@@ -222,7 +222,7 @@ arma::mat Stack::trainingData(Layer* pThatLayer)
 	return thisBatch;
 }
 
-size_t Stack::loadTraining()
+size_t Stack::loadTraining(bool doNormalize)
 {
 	uint32_t numTraining = 0;
 	uint32_t numVisible = 0;
@@ -263,6 +263,11 @@ size_t Stack::loadTraining()
 	}
 	fclose(pFile);
 	std::cout << "Loaded " << numTraining << " training samples\n";
+	
+	if (doNormalize)
+	{
+		m_trainingData = Rbm::normalize(m_trainingData);
+	}
 	return numTraining;
 }
 

source/Stack.hpp

@@ -54,7 +54,7 @@ public:
 	size_t numTraining();
 	void addTraining(const arma::mat &toAdd);
 	void delTraining(int index);
-	size_t loadTraining();
+	size_t loadTraining(bool doNormalize=false);
 	size_t saveTraining();
 	arma::mat& trainingData();
 	arma::mat trainingData(Layer *pLayer);