Updated examples to work with new ridge regression interface.

examples/krr_classification_ex.cpp

@@ -85,9 +85,11 @@ int main()
     krr_trainer<kernel_type> trainer;
 
     // The krr_trainer has the ability to perform leave-one-out cross-validation.
-    // This function tells it to measure errors in terms of the number of classification
+    // It does this to automatically determine the regularization parameter.  Since
-    // mistakes instead of mean squared error between decision function output values
+    // we are performing classification instead of regression we should be sure to
-    // and labels.  Which is what we want to do since we are performing classification.
+    // call use_classification_loss_for_loo_cv().  This function tells it to measure 
+    // errors in terms of the number of classification mistakes instead of mean squared 
+    // error between decision function output values and labels.  
     trainer.use_classification_loss_for_loo_cv();
 
 
@@ -98,11 +100,14 @@ int main()
         // tell the trainer the parameters we want to use
         trainer.set_kernel(kernel_type(gamma));
 
-        double loo_error;
+        // loo_values will contain the LOO predictions for each sample.  In the case
-        trainer.train(samples, labels, loo_error);
+        // of perfect prediction it will end up being a copy of labels.
+        std::vector<double> loo_values; 
+        trainer.train(samples, labels, loo_values);
 
-        // Print gamma and the fraction of samples misclassified during LOO cross-validation.
+        // Print gamma and the fraction of samples correctly classified during LOO cross-validation.
-        cout << "gamma: " << gamma << "     LOO error: " << loo_error << endl;
+        const double classification_accuracy = mean_sign_agreement(labels, loo_values);
+        cout << "gamma: " << gamma << "     LOO accuracy: " << classification_accuracy << endl;
     }
 
 

examples/krr_regression_ex.cpp

@@ -78,14 +78,16 @@ int main()
     // column is the output from the krr estimate.  
 
 
-    // Note that the krr_trainer has the ability to tell us the leave-one-out cross-validation
+    // Note that the krr_trainer has the ability to tell us the leave-one-out predictions
-    // accuracy.  The train() function has an optional 3rd argument and if we give it a double
+    // for each sample.  
-    // it will give us back the LOO error.
+    std::vector<double> loo_values;
-    double loo_error;
+    trainer.train(samples, labels, loo_values);
-    trainer.train(samples, labels, loo_error);
+    cout << "mean squared LOO error: " << mean_squared_error(labels,loo_values) << endl;
-    cout << "mean squared LOO error: " << loo_error << endl;
+    cout << "R^2 LOO value:          " << r_squared(labels,loo_values) << endl;
     // Which outputs the following:
-    // mean squared LOO error: 8.29563e-07
+    // mean squared LOO error: 8.29575e-07
+    // R^2 LOO value:          0.999995
+