def test_polynomial_kernel(sample_dim, num_samples,
poly_degree, poly_intercept):
"""Tests specific for Polynomial kernel."""
poly = PolyKernel(degree=poly_degree, b=poly_intercept, skip_input_checks=False)
_test_for_all_kernels(poly, sample_dim)
_test_func_is_valid_kernel(poly, sample_dim, num_samples)
DEFINED_KERNEL_FUNCS
from kernelmethods import KernelMatrix, KMAccessError, KernelMethodsException
from kernelmethods.base import ConstantKernelMatrix
from kernelmethods.operations import is_PSD
num_samples = np.random.randint(30, 100)
sample_dim = np.random.randint(3, 10) # 2
target_label_set = [1, 2]
num_samples_two = np.random.randint(30, 100)
sample_two_dim = sample_dim
sample_data = np.random.rand(num_samples, sample_dim)
target_labels = np.random.choice(target_label_set, num_samples)
poly = PolyKernel(degree=2, skip_input_checks=True)
# suffix 1 to indicate one sample case
km1 = KernelMatrix(poly)
km1.attach_to(sample_data)
max_num_elements = max_num_ker_eval = num_samples * (num_samples + 1) / 2
def test_symmetry():
if not np.isclose(km1.full, km1.full.T).all():
print('KM not symmetric')
def test_PSD():
BaseKernelFunction
from kernelmethods.numeric_kernels import GaussianKernel, LinearKernel, PolyKernel
from kernelmethods.sampling import make_kernel_bucket
num_samples = 50 # 9
sample_dim = 3 # 2
target_label_set = [1, 2]
sample_data = np.random.rand(num_samples, sample_dim)
target_labels = np.random.choice(target_label_set, (num_samples, 1))
IdealKM = target_labels.dot(target_labels.T)
rbf = KernelMatrix(GaussianKernel(sigma=10, skip_input_checks=True))
lin = KernelMatrix(LinearKernel(skip_input_checks=True))
poly = KernelMatrix(PolyKernel(degree=2, skip_input_checks=True))
# lin.attach_to(sample_data)
# rbf.attach_to(sample_data)
# poly.attach_to(sample_data)
kset = KernelSet([lin, poly, rbf])
print(kset)
def test_creation():
try:
ks = KernelSet()
except:
raise SyntaxError('empty set creation failed.')
LinearKernel, PolyKernel)
from kernelmethods.sampling import make_kernel_bucket
from kernelmethods.tests.test_numeric_kernels import _test_for_all_kernels
default_feature_dim = 10
range_feature_dim = [10, 500]
range_num_samples = [50, 500]
num_samples = np.random.randint(20)
sample_dim = np.random.randint(10)
range_polynomial_degree = [2, 10] # degree=1 is tested in LinearKernel()
np.random.seed(42)
# choosing skip_input_checks=False will speed up test runs
# default values for parameters
SupportedKernels = (GaussianKernel(), PolyKernel(), LinearKernel(),
LaplacianKernel())
num_tests_psd_kernel = 3
def gen_random_array(dim):
"""To better control precision and type of floats"""
# TODO input sparse arrays for test
return np.random.rand(dim)
def gen_random_sample(num_samples, sample_dim):
"""To better control precision and type of floats"""
# TODO input sparse arrays for test