error = SVM.test(parameters, test_data, test_labels)
print("Test error: " + str(error))
# kernels
def normal_kernel(x, y):
return np.dot(x, y)
def gaussian_kernel(x, y, gamma):
return np.exp(-np.sum(np.square(x - y)) / gamma)
# # Question 2.3.A
print("\n#####SVM in dual domain.######")
kernel = lambda x, y: normal_kernel(x, y)
for C in Cs:
print('\nC value = ' + str(C))
parameters = SVM.DualSVM(kernel, C, train_data, train_labels,
test_data, test_labels)
print(parameters)
# error = SVM.test(parameters, train_data, train_labels)
# print("Train error: " + str(error))
# error = SVM.test(parameters, test_data, test_labels)
# print("Test error: " + str(error))
# Question 2.3.B
print("\n ####SVM in dual domain with Gaussian kernel#######")
gammas = [0.1, 0.5, 1, 5, 100]
for C in Cs:
print('\nC value = ' + str(C))
for gamma in gammas:
print('Gamma value = ' + str(gamma))
kernel = lambda x, y: gaussian_kernel(x, y, gamma)
parameters = SVM.DualSVM(kernel, C, train_data, train_labels,
