Implement grid search and determine best parameters for C and sigma

ex6/dataset3Params.m

@@ -2,8 +2,8 @@ function [C, sigma] = dataset3Params(X, y, Xval, yval)
 %EX6PARAMS returns your choice of C and sigma for Part 3 of the exercise
 %where you select the optimal (C, sigma) learning parameters to use for SVM
 %with RBF kernel
-%   [C, sigma] = EX6PARAMS(X, y, Xval, yval) returns your choice of C and 
+%   [C, sigma] = EX6PARAMS(X, y, Xval, yval) returns your choice of C and
-%   sigma. You should complete this function to return the optimal C and 
+%   sigma. You should complete this function to return the optimal C and
 %   sigma based on a cross-validation set.
 %
 
@@ -15,19 +15,46 @@ sigma = 0.3;
 % Instructions: Fill in this function to return the optimal C and sigma
 %               learning parameters found using the cross validation set.
 %               You can use svmPredict to predict the labels on the cross
-%               validation set. For example, 
+%               validation set. For example,
 %                   predictions = svmPredict(model, Xval);
 %               will return the predictions on the cross validation set.
 %
-%  Note: You can compute the prediction error using 
+%  Note: You can compute the prediction error using
 %        mean(double(predictions ~= yval))
 %
 
-
+grid_search = 0;
-
+
-
+if grid_search
-
+  % Grid search
-
+  load ex6data3.mat
+  %grid = [0.01, 0.03];
+  grid = [0.01, 0.03, 0.1, 0.3, 1, 3, 10, 30];
+  results = [];
+  for C = grid
+    for sigma = grid
+
+      fprintf('== C = %.2f, sigma = %.2f\n', C, sigma);
+      model= svmTrain(X, y, C, @(x1, x2) gaussianKernel(x1, x2, sigma));
+      predictions = svmPredict(model, Xval);
+      error = mean(double(predictions ~= yval));
+      fprintf('error = %.2f\n\n', error);
+
+      results(end + 1,:) = [C, sigma, error];
+
+    end
+  end
+
+  [_, best_i] = min(results(:,3));
+  C = results(best_i, 1);
+  sigma = results(best_i, 2);
+  error = results(best_i, 3);
+  fprintf('Best: C = %.2f, sigma = %.2f with error = %.2f\n', C, sigma, error);
+else
+  % Found through the grid search above
+  C = 1.00;
+  sigma = 0.10;
+end
 
 % =========================================================================