# SPDX-License-Identifier: MPL-2.0
# Copyright (C) 2020- The University of Tokyo
#
# This Source Code Form is subject to the terms of the Mozilla Public
# License, v. 2.0. If a copy of the MPL was not distributed with this
# file, You can obtain one at https://mozilla.org/MPL/2.0/.
import numpy as np
[ドキュメント]
class gauss:
"""Gaussian likelihood function"""
def __init__(self, std=1, max_params=1e6, min_params=1e-6):
"""
Parameters
----------
std: numpy.ndarray or float
standard deviation.
max_params: float
The maximum value of the parameter.
If the parameter is greater than this value, it will be replaced by this value.
min_params: float
The minimum value of the parameter.
If the parameter is less than this value, it will be replaced by this value.
"""
self.min_params = np.log(min_params)
self.max_params = np.log(max_params)
self.num_params = 1
self.std = std
self.params = np.log(std)
self.set_params(self.params)
[ドキュメント]
def supp_params(self, params=None):
"""
Set maximum (minimum) values for parameters when the parameter is greater(less) than this value.
Parameters
----------
params: numpy.ndarray
Parameters for optimization.
Array of real elements of size (n,), where ‘n’ is the number of independent variables.
Returns
-------
"""
if params is None:
params = np.copy(params)
if params > self.max_params:
params = self.max_params
if params < self.min_params:
params = self.min_params
return params
[ドキュメント]
def trans_params(self, params=None):
"""
Get exp[params].
Parameters
----------
params: numpy.ndarray
Parameters for optimization.
Array of real elements of size (n,), where ‘n’ is the number of independent variables.
Returns
-------
std: numpy.ndarray
"""
if params is None:
params = np.copy(self.params)
std = np.exp(params)
return std
[ドキュメント]
def get_params_bound(self):
"""
Get boundary array.
Returns
-------
bound: list
A num_params-dimensional array with the tuple (min_params, max_params).
"""
bound = [(self.min_params, self.max_params) for i in range(0, self.num_params)]
return bound
[ドキュメント]
def get_cov(self, num_data, params=None):
"""
Get a covariance matrix
Parameters
----------
num_data: int
params: numpy.ndarray
Parameters for optimization.
Array of real elements of size (n,), where ‘n’ is the number of independent variables.
Returns
-------
numpy.ndarray
Diagonal element matrix of exp(2.0*params)
"""
std = self.trans_params(params)
var = std**2
return var * np.identity(num_data)
[ドキュメント]
def get_grad(self, num_data, params=None):
"""
Get a gradient matrix
Parameters
----------
num_data: int
params: numpy.ndarray
Parameters for optimization.
Array of real elements of size (n,), where ‘n’ is the number of independent variables.
Returns
-------
numpy.ndarray
Diagonal element matrix of 2.0 * exp(2.0*params)
"""
std = self.trans_params(params)
var = std**2
return var * np.identity(num_data) * 2
[ドキュメント]
def set_params(self, params):
"""
Set parameters.
Parameters
----------
params: numpy.ndarray
Parameters for optimization.
Array of real elements of size (n,), where ‘n’ is the number of independent variables.
Returns
-------
"""
self.params = self.supp_params(params)
self.std = self.trans_params(params)
[ドキュメント]
def get_cand_params(self, t):
"""
Getting candidate parameters.
Parameters
----------
t: numpy.ndarray
N dimensional array. The negative energy of each search candidate (value of the objective function to be optimized).
Returns
-------
numpy.ndarray
log[ standard deviation of t] - log 10.0
"""
return np.log(np.std(t) / 10)
# [TODO] Check: This function seems not to be used.
[ドキュメント]
def sampling(self, fmean):
"""
Sampling by adding noise
Parameters
----------
fmean: numpy.ndarray
Returns
-------
"""
num_data = fmean.shape[0]
eps = self.std * np.random.randn(num_data)
return fmean + eps