X_train_features = np.column_stack((X_train_features_1_ltr_imgs,X_train_features_1_rtl_imgs,X_train_features_1_ttb_imgs,X_train_features_1_btt_imgs,X_train_features_0_ltr_imgs,X_train_features_0_rtl_imgs,X_train_features_0_btt_imgs,X_train_features_0_ttb_imgs)) X_test_features = np.column_stack((X_test_features_1_ltr_imgs,X_test_features_1_rtl_imgs,X_test_features_1_ttb_imgs,X_test_features_1_btt_imgs,X_test_features_0_ltr_imgs,X_test_features_0_rtl_imgs,X_test_features_0_btt_imgs,X_test_features_0_ttb_imgs)) ridge_model = RidgeClassifier().fit(X_train_features, y_train) images_train_accuracy_ridge[t] = ridge_model.score(X_train_features, y_train) images_test_accuracy_ridge[t] = ridge_model.score(X_test_features, y_test) c = 10 svm_model = SVC(kernel='rbf', C = c).fit(X_train_features, y_train) images_train_accuracy_svm[t] = svm_model.score(X_train_features, y_train) images_test_accuracy_svm[t] = svm_model.score(X_test_features, y_test) ### Landscape Features i = 3 j = 50 X_train_features_1_ltr_landscapes, X_test_features_1_ltr_landscapes = landscape_features(one_dim_ltr_train, one_dim_ltr_test, num_landscapes=i, resolution=j) X_train_features_0_ltr_landscapes, X_test_features_0_ltr_landscapes = landscape_features(zero_dim_ltr_train, zero_dim_ltr_test, num_landscapes=i, resolution=j) X_train_features_1_rtl_landscapes, X_test_features_1_rtl_landscapes = landscape_features(one_dim_rtl_train, one_dim_rtl_test, num_landscapes=i, resolution=j) X_train_features_0_rtl_landscapes, X_test_features_0_rtl_landscapes = landscape_features(zero_dim_rtl_train, zero_dim_rtl_test, num_landscapes=i, resolution=j) X_train_features_1_ttb_landscapes, X_test_features_1_ttb_landscapes = landscape_features(one_dim_ttb_train, one_dim_ttb_test, num_landscapes=i, resolution=j) X_train_features_0_ttb_landscapes, X_test_features_0_ttb_landscapes = landscape_features(zero_dim_ttb_train, zero_dim_ttb_test, num_landscapes=i, resolution=j) X_train_features_1_btt_landscapes, X_test_features_1_btt_landscapes = landscape_features(one_dim_btt_train, one_dim_btt_test, num_landscapes=i, resolution=j) X_train_features_0_btt_landscapes, X_test_features_0_btt_landscapes = landscape_features(zero_dim_btt_train, zero_dim_btt_test, num_landscapes=i, resolution=j) X_train_features = np.column_stack((X_train_features_1_ltr_landscapes,X_train_features_1_rtl_landscapes,X_train_features_1_ttb_landscapes,X_train_features_1_btt_landscapes,X_train_features_0_ltr_landscapes,X_train_features_0_rtl_landscapes,X_train_features_0_btt_landscapes,X_train_features_0_ttb_landscapes)) X_test_features = np.column_stack((X_test_features_1_ltr_landscapes,X_test_features_1_rtl_landscapes,X_test_features_1_ttb_landscapes,X_test_features_1_btt_landscapes,X_test_features_0_ltr_landscapes,X_test_features_0_rtl_landscapes,X_test_features_0_btt_landscapes,X_test_features_0_ttb_landscapes)) ridge_model = RidgeClassifier().fit(X_train_features, y_train)
X_test_features_X1_imgs, X_test_features_Y1_imgs, X_test_features_Z1_imgs, X_test_features_H1_imgs, X_test_features_S1_imgs, X_test_features_V1_imgs)) ridge_model = RidgeClassifier().fit(X_train_features, y_train) images_train_accuracy_ridge[k] = ridge_model.score(X_train_features, y_train) images_test_accuracy_ridge[k] = ridge_model.score(X_test_features, y_test) svm_model = SVC(kernel='rbf', C=1).fit(X_train_features, y_train) images_train_accuracy_svm[k] = svm_model.score(X_train_features, y_train) images_test_accuracy_svm[k] = svm_model.score(X_test_features, y_test) ### Landscapes i = 10 j = 50 X_train_features_R0_landscapes, X_test_features_R0_landscapes = landscape_features( R0_train_sample, R0_test_sample, num_landscapes=i, resolution=j) X_train_features_G0_landscapes, X_test_features_G0_landscapes = landscape_features( G0_train_sample, G0_test_sample, num_landscapes=i, resolution=j) X_train_features_B0_landscapes, X_test_features_B0_landscapes = landscape_features( B0_train_sample, B0_test_sample, num_landscapes=i, resolution=j) X_train_features_X0_landscapes, X_test_features_X0_landscapes = landscape_features( X0_train_sample, X0_test_sample, num_landscapes=i, resolution=j) X_train_features_Y0_landscapes, X_test_features_Y0_landscapes = landscape_features( Y0_train_sample, Y0_test_sample, num_landscapes=i, resolution=j) X_train_features_Z0_landscapes, X_test_features_Z0_landscapes = landscape_features( Z0_train_sample, Z0_test_sample, num_landscapes=i, resolution=j) X_train_features_H0_landscapes, X_test_features_H0_landscapes = landscape_features( H0_train_sample, H0_test_sample, num_landscapes=i, resolution=j) X_train_features_S0_landscapes, X_test_features_S0_landscapes = landscape_features( S0_train_sample, S0_test_sample, num_landscapes=i, resolution=j) X_train_features_V0_landscapes, X_test_features_V0_landscapes = landscape_features(
### Kernel Features s = .4 X_train_features_1_kernel, X_test_features_1_kernel = fast_kernel_features( X_dgm1_train, X_dgm1_test, s) X_train_features_0_kernel, X_test_features_0_kernel = fast_kernel_features( X_dgm0_train, X_dgm0_test, s) X_train_features = X_train_features_1_kernel + X_train_features_0_kernel X_test_features = X_test_features_1_kernel + X_test_features_0_kernel svm_model = SVC(kernel='precomputed').fit(X_train_features, y_train) kernel_train_accuracy_svm[k] = svm_model.score(X_train_features, y_train) kernel_test_accuracy_svm[k] = svm_model.score(X_test_features, y_test) ### Landscape Features n = 5 r = 100 X_train_features_1_landscapes, X_test_features_1_landscapes = landscape_features( X_dgm1_train, X_dgm1_test, num_landscapes=n, resolution=r) X_train_features_0_landscapes, X_test_features_0_landscapes = landscape_features( X_dgm0_train, X_dgm0_test, num_landscapes=n, resolution=r) X_train_features = np.column_stack( (X_train_features_0_landscapes, X_train_features_1_landscapes)) X_test_features = np.column_stack( (X_test_features_0_landscapes, X_test_features_1_landscapes)) ridge_model = RidgeClassifier().fit(X_train_features, y_train) landscapes_train_accuracy_ridge[k] = ridge_model.score( X_train_features, y_train) landscapes_test_accuracy_ridge[k] = ridge_model.score( X_test_features, y_test) c = 1 svm_model = SVC(kernel='rbf', C=c).fit(X_train_features, y_train)