def speech_fourier_3be(ker_sigma, regu, upper_freqs, domain=np.array([[0,
1]])):
# Fourier output basis
output_basis_params = {
"lower_freq": 0,
"upper_freq": upper_freqs,
"domain": domain
}
output_bases = configs_generation.subconfigs_combinations(
"fourier", output_basis_params, exclude_list=["domain"])
# Sum of Gaussian kernels
ker_sigmas = ker_sigma * np.ones(13)
gauss_kers = [
kernels.GaussianScalarKernel(sig, normalize=False, normalize_dist=True)
for sig in ker_sigmas
]
multi_ker = kernels.SumOfScalarKernel(gauss_kers, normalize=False)
# Generate full configs
params = {
"kernel": multi_ker,
"basis_out": output_bases,
"regu": regu,
"center_output": True
}
configs = configs_generation.configs_combinations(params)
# Create list of regressors from that config
regs = [triple_basis.BiBasisEstimator(**config) for config in configs]
return configs, regs
def speech_fpca_3be(ker_sigma,
regu,
n_fpca,
n_evals_fpca,
domain=np.array([[0, 1]])):
# FPCA output basis
output_basis_params = {
"n_basis": n_fpca,
"input_dim": 1,
"domain": domain,
"n_evals": n_evals_fpca
}
output_bases = configs_generation.subconfigs_combinations(
"functional_pca", output_basis_params, exclude_list=["domain"])
# Sum of Gaussian kernels
ker_sigmas = ker_sigma * np.ones(13)
gauss_kers = [
kernels.GaussianScalarKernel(sig, normalize=False, normalize_dist=True)
for sig in ker_sigmas
]
multi_ker = kernels.SumOfScalarKernel(gauss_kers, normalize=False)
# Generate full configs
params = {
"kernel": multi_ker,
"basis_out": output_bases,
"regu": regu,
"center_output": True
}
configs = configs_generation.configs_combinations(params)
# Create list of regressors from that config
regs = [triple_basis.BiBasisEstimator(**config) for config in configs]
return configs, regs
def kernel_generator_ke_speech(kx_sigma):
if isinstance(kx_sigma, Iterable):
multi_sigs = [sig * np.ones(13) for sig in kx_sigma]
bases_kers = [[
kernels.GaussianScalarKernel(sig,
normalize=False,
normalize_dist=True)
for sig in multi_sig
] for multi_sig in multi_sigs]
kxs = [
kernels.SumOfScalarKernel(base_ker, normalize=False)
for base_ker in bases_kers
]
else:
multi_sig = kx_sigma * np.ones(13)
base_ker = [
kernels.GaussianScalarKernel(sig,
normalize=False,
normalize_dist=True)
for sig in multi_sig
]
kxs = kernels.SumOfScalarKernel(base_ker, normalize=False)
return kxs
def kernels_generator_gauss_fkrr_speech(kin_sigma, kout_sigma):
kin_sigmas = kin_sigma * np.ones(13)
gauss_kers = [
kernels.GaussianScalarKernel(sig, normalize=False, normalize_dist=True)
for sig in kin_sigmas
]
kernels_in = kernels.SumOfScalarKernel(gauss_kers, normalize=False)
if isinstance(kout_sigma, Iterable):
kernels_out = [
kernels.GaussianScalarKernel(sig, normalize=False)
for sig in kout_sigma
]
else:
kernels_out = kernels.GaussianScalarKernel(kout_sigma, normalize=False)
return kernels_in, kernels_out
def speech_rffs_kpl(kernel_sigma,
regu,
n_rffs,
rffs_sigma,
seed_rffs,
center_output,
domain=np.array([[0, 1]])):
# FPCA output basis
output_basis_params = {
"n_basis": n_rffs,
"bandwidth": rffs_sigma,
"input_dim": 1,
"domain": domain,
"seed": seed_rffs
}
output_bases = configs_generation.subconfigs_combinations(
"random_fourier", output_basis_params, exclude_list=["domain"])
# Sum of Gaussian kernels
kernel_sigmas = kernel_sigma * np.ones(13)
gauss_kers = [
kernels.GaussianScalarKernel(sig, normalize=False, normalize_dist=True)
for sig in kernel_sigmas
]
multi_ker = kernels.SumOfScalarKernel(gauss_kers, normalize=False)
# Penalize power
output_matrix_params = {}
output_matrices = configs_generation.subconfigs_combinations(
"eye", output_matrix_params)
# Generate full configs
params = {
"kernel": multi_ker,
"B": output_matrices,
"basis_out": output_bases,
"regu": regu,
"center_output": center_output
}
configs = configs_generation.configs_combinations(params)
# Create list of regressors from that config
regs = [kproj_learning.SeperableKPL(**config) for config in configs]
return configs, regs
def speech_fpca_penpow_kpl(kernel_sigma,
regu,
n_fpca,
n_evals_fpca,
decrease_base,
domain=np.array([[0, 1]])):
# FPCA output basis
output_basis_params = {
"n_basis": n_fpca,
"input_dim": 1,
"domain": domain,
"n_evals": n_evals_fpca
}
output_bases = configs_generation.subconfigs_combinations(
"functional_pca", output_basis_params, exclude_list=["domain"])
# Sum of Gaussian kernels
kernel_sigmas = kernel_sigma * np.ones(13)
gauss_kers = [
kernels.GaussianScalarKernel(sig, normalize=False, normalize_dist=True)
for sig in kernel_sigmas
]
multi_ker = kernels.SumOfScalarKernel(gauss_kers, normalize=False)
# Penalize power
output_matrix_params = {"decrease_base": decrease_base}
output_matrices = configs_generation.subconfigs_combinations(
"pow", output_matrix_params)
# Generate full configs
params = {
"kernel": multi_ker,
"B": output_matrices,
"basis_out": output_bases,
"regu": regu,
"center_output": True
}
configs = configs_generation.configs_combinations(params)
# Create list of regressors from that config
regs = [kproj_learning.SeperableKPL(**config) for config in configs]
return configs, regs
