Adding some suggestion

src/mlpack/methods/ann/activation_functions/gaussian_function.hpp

@@ -2,7 +2,7 @@
  * @file gaussian_function.hpp
  * @author Himanshu Pathak
  *
- * Definition and implementation of the logistic function.
+ * Definition and implementation of the gaussian function.
  *
  * mlpack is free software; you may redistribute it and/or modify it under the
  * terms of the 3-clause BSD license.  You should have received a copy of the
@@ -18,12 +18,11 @@ namespace mlpack {
 namespace ann /** Artificial Neural Network. */ {
 
 /**
- * The logistic function, defined by
+ * The gaussian function, defined by
  *
  * @f{eqnarray*}{
- * f(x) &=& \frac{1}{1 + e^{-x}} \\
- * f'(x) &=& f(x) * (1 - f(x)) \\
- * f^{-1}(y) &=& ln(\frac{y}{1-y})
+ * f(x) &=& \frac{1}{e^{-1 * x^2}} \\
+ * f'(x) &=& f(x) * -x * 2)
  * @f}
  */
 class GaussianFunction
@@ -54,7 +53,7 @@ class GaussianFunction
   }
 
   /**
-   * Computes the first derivative of the logistic function.
+   * Computes the first derivative of the gaussian function.
    *
    * @param x Input data.
    * @return f'(x)
@@ -65,7 +64,7 @@ class GaussianFunction
   }
 
   /**
-   * Computes the first derivatives of the logistic function.
+   * Computes the first derivatives of the gaussian function.
    *
    * @param y Input activations.
    * @param x The resulting derivatives.

src/mlpack/methods/ann/layer/radial_basis_function.hpp

@@ -2,8 +2,7 @@
  * @file radial_basis_function.hpp
  * @author Hiamsnshu Pathak
  *
- * Definition of the Dropout class, which implements a regularizer that
- * randomly sets units to zero preventing units from co-adapting.
+ *
  *
  * mlpack is free software; you may redistribute it and/or modify it under the
  * terms of the 3-clause BSD license.  You should have received a copy of the
@@ -31,16 +30,18 @@ template<typename InputDataType = arma::mat,
 class RBF
 {
  public:
+  //! Create the Linear object.
   RBF();
- /**
+
+  /**
    * Create the Radial Basis Function layer object using the specified
    * parameters.
    *
    * @param inSize The number of input units.
    * @param outSize The number of output units.
    */
   RBF(const size_t inSize,
-    const size_t outSize);
+      const size_t outSize);
 
   /**
    * Ordinary feed forward pass of the radial basis function.

src/mlpack/methods/ann/layer/radial_basis_function_impl.hpp

@@ -43,26 +43,25 @@ void RBF<InputDataType, OutputDataType>::Forward(
 {
   centres = arma::mat(outSize, input.n_cols, arma::fill::randu);
   centres = arma::normcdf(centres, 0, 1);
-  sigmas = arma::ones(1,outSize);
+  sigmas = arma::ones(1, outSize);
   arma::cube x = arma::cube(outSize, input.n_cols, input.n_rows);
-  
-  for(size_t i=0;i < input.n_rows;i++)
+
+  for (size_t i = 0; i < input.n_rows; i++)
   {
-    for (size_t j = 0;j < outSize; j++)
-    {
-      x.slice(i).row(j) = input.row(i);
-    }
+    x.slice(i).each_row() = input.row(i);
   }
+
   arma::cube c = arma::cube(outSize, input.n_cols, input.n_rows);
-  for(size_t i=0; i < input.n_rows; i++)
-  {
-    c.slice(i)= centres;
-  }
+  c.each_slice()= centres;
+
   distances = arma::mat(input.n_rows, outSize);
-
-  for(size_t i=0;i < outSize;i++)
-  { 
-  distances.row(i) = arma::pow(arma::sum (arma::pow ((x.slice(i) - c.slice(i)), 2), 1), 0.5).t() * sigmas(i);
+
+  for (size_t i = 0; i < outSize; i++)
+  {
+  distances.row(i) = arma::pow (arma::sum (
+                                arma::pow ((
+                                x.slice(i) - c.slice(i)),
+                                2), 1), 0.5).t() * sigmas(i);
   }
 
   output = distances;

src/mlpack/tests/feedforward_network_test.cpp

@@ -649,7 +649,6 @@ BOOST_AUTO_TEST_CASE(RBFNetworkTest)
    * |     |
    * +-----+
    */
-
 
   FFN<NegativeLogLikelihood<> > model;
   model.Add<RBF<> >(trainData.n_cols, 8);
@@ -659,31 +658,6 @@ BOOST_AUTO_TEST_CASE(RBFNetworkTest)
   model.Add<LogSoftMax<> >();
   std::cout<<trainData.n_cols;
 
-  // Vanilla neural net with logistic activation function.
-  // Because 92% of the patients are not hyperthyroid the neural
-  // network must be significant better than 92%.
-  TestNetwork<>(model, trainData, trainLabels, testData, testLabels, 10, 0.1);
-  arma::mat dataset;
-  dataset.load("mnist_first250_training_4s_and_9s.arm");
-
-  // Normalize each point since these are images.
-  for (size_t i = 0; i < dataset.n_cols; ++i)
-  {
-    dataset.col(i) /= norm(dataset.col(i), 2);
-  }
-
-  arma::mat labels = arma::zeros(1, dataset.n_cols);
-  labels.submat(0, labels.n_cols / 2, 0, labels.n_cols - 1).fill(1);
-  labels += 1;
-
-  FFN<NegativeLogLikelihood<> > model1;
-  model1.Add<RBF<> >(dataset.n_cols, 10);
-  model1.Add<GaussianFunctionLayer<> >();
-  model1.Add<Linear<> >(dataset.n_rows, 10);
-  model.Add<Linear<> >(10, 2);
-  model1.Add<LogSoftMax<> >();
-  // Vanilla neural net with logistic activation function.
-  TestNetwork<>(model1, dataset, labels, dataset, labels, 10, 0.2);
 }
 
 BOOST_AUTO_TEST_SUITE_END();