correlated_parameter_distributions

circuit_analyzer.all.correlated_parameter_distributions(names, means, cov, sample_size, seed=None)

Generate a dictionary of correlated normal distributions.

Parameters:

nameslist

list of names of the parameters that you want to sample. If a parameter is a list, tuple or an array, for instance ‘name = [1, 2, 4]’, then split that parameter into ‘name:0’, ‘name:1’ and ‘name:2’.

meanslist

list of the means of the parameters that you want to sample. Same length of ‘names’.

cov: np.ndarray

covariance matrix

sample_sizeint

number of samples

seedNone or int or numpy.random.RandomState instance, optional

This parameter defines the RandomState object to use for drawing random variates. If None (or np.random), the global np.random state is used. If integer, it is used to seed the local RandomState instance. Default is None.

Returns:

parameter_distributiondict

a dict with per parameter name a normal distribution.

Examples

>>> import circuit_analyzer.all as ca
>>> import numpy as np
>>> var_n_eff = 0.01**2
>>> var_n_g = 0.01**2
>>> var_loss_dB_cm = 0.1**2
>>> var_sc_1 = (0.5**0.5 * 0.01) ** 2
>>> var_sc_2 = (0.5**0.5 * 0.01) ** 2
>>> covar1 = 0.5 * np.sqrt(var_n_eff * var_n_g)
>>> covar2 = np.sqrt(var_sc_1 * var_sc_2)
>>> parameter_distribution = ca.correlated_parameter_distributions(
>>>     names=["n_eff", "n_g", "loss_dB_cm", "straight_coupling1", "straight_coupling2"],
>>>     means=[1.0, 2.0, 2.0, 0.5**0.5, 0.5**0.5],
>>>     cov=np.array(
>>>         [
>>>             [var_n_eff, covar1, 0.0, 0.0, 0.0],
>>>             [covar1, var_n_g, 0.0, 0.0, 0.0],
>>>             [0.0, 0.0, var_loss_dB_m, 0.0, 0.0],
>>>             [0.0, 0.0, 0.0, var_sc_1, covar2],
>>>             [0.0, 0.0, 0.0, covar2, var_sc_2],
>>>         ]
>>>     ),
>>>     sample_size=80,
>>>     seed=0,
>>> )