- use normalize per training sample

git-svn-id: http://moon:8086/svn/software/trunk/projects/RBM@555 b431acfa-c32f-4a4a-93f1-934dc6c82436

Source/Rbm.cpp

@@ -66,7 +66,7 @@ MatrixXd Rbm::sample(MatrixXd const &src)
 
 	noiseUniform(n);
 
-	return (src.array() > n.array()).cast<double>();
+	return (src.array() >= n.array()).cast<double>();
 }
 
 void Rbm::sample(MatrixXd &dst, MatrixXd const &src)
@@ -86,6 +86,7 @@ RowVectorXd Rbm::probsLogistic(RowVectorXd const &src)
 
 MatrixXd Rbm::normalizeData(MatrixXd const &src)
 {
+#if 0
 	double mean = src.array().mean();
 	cout << "mean" << ": " << endl << mean << endl;
 
@@ -94,7 +95,19 @@ MatrixXd Rbm::normalizeData(MatrixXd const &src)
 	double stddev = sqrt(x2.array().mean());
 	cout << "stddev" << ": " << endl << stddev << endl;
 	
-	return x;
+	return x/stddev;
+	
+#else
+	MatrixXd mean = src.colwise().mean();
+	MatrixXd x = src - mean.replicate(src.rows(), 1);
+	MatrixXd x2 = x.cwiseProduct(x);
+	
+	
+	MatrixXd stddev_inv = x2.colwise().mean();
+	stddev_inv = stddev_inv.array().sqrt().cwiseInverse();
+	return x*stddev_inv.replicate(src.rows(), 1);
+	
+#endif
 	
 }