cosmetic and comment changes

Author: dclambert

elm.py

@@ -245,7 +245,7 @@ class GenELMClassifier(BaseELM, ClassifierMixin):
     `classes_` : numpy array of shape [n_classes]
         Array of class labels
 
-    `elm_regressor_` : ELMRegressor instance
+    `genelm_regressor_` : ELMRegressor instance
         Performs actual fit of binarized values
 
     See Also
@@ -344,8 +344,8 @@ class ELMRegressor(BaseEstimator, RegressorMixin):
     [1][2]
 
     ELMRegressor is a wrapper for an GenELMRegressor that uses a
-    RandomLayer and exposes the RandomLayer's parameters in its
-    own constructor.
+    RandomLayer and passes the __init__ parameters through
+    to the hidden layer generated by the fit() method.
 
     Parameters
     ----------
@@ -418,9 +418,11 @@ def __init__(self, n_hidden=20, alpha=0.5, rbf_width=1.0,
         self.rbf_width = rbf_width
         self.regressor = regressor
 
-        self._genelm_regressor_ = None
+        self._genelm_regressor = None
 
     def _create_random_layer(self):
+        """Pass init params to RandomLayer"""
+
         return RandomLayer(n_hidden=self.n_hidden,
                            alpha=self.alpha, random_state=self.random_state,
                            activation_func=self.activation_func,
@@ -449,9 +451,9 @@ def fit(self, X, y):
             Returns an instance of self.
         """
         rhl = self._create_random_layer()
-        self.genelm_regressor_ = GenELMRegressor(hidden_layer=rhl,
+        self._genelm_regressor = GenELMRegressor(hidden_layer=rhl,
                                                  regressor=self.regressor)
-        self.genelm_regressor_.fit(X, y)
+        self._genelm_regressor.fit(X, y)
         return self
 
     def predict(self, X):
@@ -467,10 +469,10 @@ def predict(self, X):
         C : numpy array of shape [n_samples, n_outputs]
             Predicted values.
         """
-        if (self.genelm_regressor_ is None):
+        if (self._genelm_regressor is None):
             raise ValueError("SimpleELMRegressor not fitted")
 
-        return self.genelm_regressor_.predict(X)
+        return self._genelm_regressor.predict(X)
 
 
 class ELMClassifier(ELMRegressor):
@@ -486,8 +488,8 @@ class ELMClassifier(ELMRegressor):
     data, then uses the superclass to compute the decision function that
     is then unbinarized to yield the prediction.
 
-    The RandomLayer used for the input transform are exposed in the
-    ELMClassifier constructor.
+    The params for the RandomLayer used in the input transform are
+    exposed in the ELMClassifier constructor.
 
     Parameters
     ----------
@@ -609,5 +611,8 @@ def predict(self, X):
         return class_predictions
 
     def score(self, X, y):
+        """Force use of accuracy score since we don't inherit
+           from ClassifierMixin"""
+
         from sklearn.metrics import accuracy_score
         return accuracy_score(y, self.predict(X))