Added some tests for the new krr_trainer. I also simplified the checkers board

dataset used to test the classifiers a little so that the test runs faster.

--HG--
extra : convert_revision : svn%3Afdd8eb12-d10e-0410-9acb-85c331704f74/trunk%403760

dlib/test/svm.cpp

@@ -175,6 +175,8 @@ namespace
         krls<kernel_type> test(kernel_type(0.1),0.001);
         rvm_regression_trainer<kernel_type> rvm_test;
         rvm_test.set_kernel(test.get_kernel());
+        krr_trainer<kernel_type> krr_test;
+        krr_test.set_kernel(test.get_kernel());
         rbf_network_trainer<kernel_type> rbf_test;
         rbf_test.set_kernel(test.get_kernel());
         rbf_test.set_num_centers(13);
@@ -198,6 +200,8 @@ namespace
         print_spinner();
         decision_function<kernel_type> test3 = rbf_test.train(samples, labels);
         print_spinner();
+        decision_function<kernel_type> test4 = krr_test.train(samples, labels);
+        print_spinner();
 
         // now we output the value of the sinc function for a few test points as well as the 
         // value predicted by krls object.
@@ -215,6 +219,11 @@ namespace
         m(0) = 0.1; dlog << LDEBUG << "rbf: " << sinc(m(0)) << "   " << test3(m); DLIB_TEST(abs(sinc(m(0)) - test3(m)) < 0.01);
         m(0) = -4;  dlog << LDEBUG << "rbf: " << sinc(m(0)) << "   " << test3(m); DLIB_TEST(abs(sinc(m(0)) - test3(m)) < 0.01);
         m(0) = 5.0; dlog << LDEBUG << "rbf: " << sinc(m(0)) << "   " << test3(m); DLIB_TEST(abs(sinc(m(0)) - test3(m)) < 0.01);
+
+        m(0) = 2.5; dlog << LDEBUG << "krr: " << sinc(m(0)) << "   " << test4(m); DLIB_TEST(abs(sinc(m(0)) - test4(m)) < 0.01);
+        m(0) = 0.1; dlog << LDEBUG << "krr: " << sinc(m(0)) << "   " << test4(m); DLIB_TEST(abs(sinc(m(0)) - test4(m)) < 0.01);
+        m(0) = -4;  dlog << LDEBUG << "krr: " << sinc(m(0)) << "   " << test4(m); DLIB_TEST(abs(sinc(m(0)) - test4(m)) < 0.01);
+        m(0) = 5.0; dlog << LDEBUG << "krr: " << sinc(m(0)) << "   " << test4(m); DLIB_TEST(abs(sinc(m(0)) - test4(m)) < 0.01);
         dlog << LINFO << "   end test_regression()";
     }
 
@@ -330,7 +339,7 @@ namespace
         std::vector<matrix<double,0,1> > x_linearized;
         std::vector<scalar_type> y;
 
-        get_checkerboard_problem(x,y, 300, 3);
+        get_checkerboard_problem(x,y, 300, 2);
         const scalar_type gamma = 1;
 
         typedef radial_basis_kernel<sample_type> kernel_type;
@@ -342,6 +351,10 @@ namespace
         rvm_trainer<kernel_type> rvm_trainer;
         rvm_trainer.set_kernel(kernel_type(gamma));
 
+        krr_trainer<kernel_type> krr_trainer;
+        krr_trainer.estimate_lambda_for_classification();
+        krr_trainer.set_kernel(kernel_type(gamma));
+
         svm_pegasos<kernel_type> pegasos_trainer;
         pegasos_trainer.set_kernel(kernel_type(gamma));
         pegasos_trainer.set_lambda(0.00001);
@@ -367,6 +380,8 @@ namespace
         print_spinner();
         matrix<scalar_type> rvm_cv = cross_validate_trainer_threaded(rvm_trainer, x,y, 4, 2);
         print_spinner();
+        matrix<scalar_type> krr_cv = cross_validate_trainer_threaded(krr_trainer, x,y, 4, 2);
+        print_spinner();
         matrix<scalar_type> svm_cv = cross_validate_trainer(trainer, x,y, 4);
         print_spinner();
         matrix<scalar_type> rbf_cv = cross_validate_trainer_threaded(rbf_trainer, x,y, 10, 2);
@@ -384,6 +399,7 @@ namespace
         print_spinner();
 
         dlog << LDEBUG << "rvm cv:        " << rvm_cv;
+        dlog << LDEBUG << "krr cv:        " << krr_cv;
         dlog << LDEBUG << "svm cv:        " << svm_cv;
         dlog << LDEBUG << "rbf cv:        " << rbf_cv;
         dlog << LDEBUG << "lin cv:        " << lin_cv;
@@ -397,6 +413,7 @@ namespace
                       sum(abs(peg_cv - peg_c_cv)) << "   \n" << peg_cv << peg_c_cv  );
 
         DLIB_TEST_MSG(mean(rvm_cv) > 0.9, rvm_cv);
+        DLIB_TEST_MSG(mean(krr_cv) > 0.9, krr_cv);
         DLIB_TEST_MSG(mean(svm_cv) > 0.9, svm_cv);
         DLIB_TEST_MSG(mean(rbf_cv) > 0.9, rbf_cv);
         DLIB_TEST_MSG(mean(lin_cv) > 0.9, lin_cv);