flip.likelihood

Attributes

jax_installed
log
_available_priors
_available_inversion_methods
log_likelihood_gaussian_inverse_jit

Classes

BaseLikelihood	Helper class that provides a standard way to create an ABC using
MultivariateGaussianLikelihood	Helper class that provides a standard way to create an ABC using
MultivariateGaussianLikelihoodInterpolate1D	Helper class that provides a standard way to create an ABC using
MultivariateGaussianLikelihoodInterpolate2D	Helper class that provides a standard way to create an ABC using
Prior
GaussianPrior
PositivePrior
UniformPrior

Functions

log_likelihood_gaussian_inverse(vector, covariance_sum)
log_likelihood_gaussian_solve(vector, covariance_sum)
log_likelihood_gaussian_cholesky(vector, covariance_sum)
log_likelihood_gaussian_cholesky_inverse(vector, ...)
log_likelihood_gaussian_cholesky_regularized(vector, ...)
no_prior(x)
prior_sum(priors, x)

Module Contents

flip.likelihood.jax_installed = True

flip.likelihood.log

flip.likelihood._available_priors = ['gaussian', 'positive', 'uniform']

flip.likelihood._available_inversion_methods = ['inverse', 'solve', 'cholesky', 'cholesky_regularized', 'cholesky_inverse']

flip.likelihood.log_likelihood_gaussian_inverse(vector, covariance_sum)

flip.likelihood.log_likelihood_gaussian_solve(vector, covariance_sum)

flip.likelihood.log_likelihood_gaussian_cholesky(vector, covariance_sum)

flip.likelihood.log_likelihood_gaussian_cholesky_inverse(vector, covariance_sum)

flip.likelihood.log_likelihood_gaussian_cholesky_regularized(vector, covariance_sum)

flip.likelihood.log_likelihood_gaussian_inverse_jit

flip.likelihood.no_prior(x)

flip.likelihood.prior_sum(priors, x)

class flip.likelihood.BaseLikelihood(covariance=None, data=None, parameter_names=None, likelihood_properties={})

Bases: abc.ABC

Helper class that provides a standard way to create an ABC using inheritance.

_default_likelihood_properties

covariance = None

data = None

parameter_names = None

free_par

likelihood_properties

prior

__call__(parameter_values)

abstract _init_likelihood(*args)

classmethod init_from_covariance(covariance, data, parameter_names, likelihood_properties={}, **kwargs)

The init_from_covariance function is a class method that initializes the likelihood object from a covariance matrix.

Parameters:

• cls – Create a new instance of the class

• covariance – Compute the full matrix of the covariance

• parameter_names – Set the names of the parameters

• density – Compute the vector and its error

• density_err – Compute the vector_err

• velocity – Compute the vector and vector_err

• velocity_err – Compute the error in the vector

:param : Compute the vector

Returns:

A likelihood object

initialize_prior()

verify_covariance()

class flip.likelihood.MultivariateGaussianLikelihood(covariance=None, data=None, parameter_names=None, likelihood_properties={})

Bases: BaseLikelihood

Helper class that provides a standard way to create an ABC using inheritance.

_init_likelihood()

class flip.likelihood.MultivariateGaussianLikelihoodInterpolate1D(covariance=None, data=None, parameter_names=None, likelihood_properties={}, interpolation_value_name=None, interpolation_value_range=None)

Bases: BaseLikelihood

Helper class that provides a standard way to create an ABC using inheritance.

interpolation_value_name = None

interpolation_value_range = None

free_par

_init_likelihood()

class flip.likelihood.MultivariateGaussianLikelihoodInterpolate2D(covariance=None, data=None, parameter_names=None, prior=None, likelihood_properties={}, interpolation_value_name_0=None, interpolation_value_name_1=None, interpolation_value_range_0=None, interpolation_value_range_1=None)

Bases: BaseLikelihood

Helper class that provides a standard way to create an ABC using inheritance.

interpolation_value_name_0 = None

interpolation_value_name_1 = None

interpolation_value_range_0 = None

interpolation_value_range_1 = None

__call__(parameter_values)

The __call__ function is the function that will be called when the likelihood object is called. It takes in a list of parameter values, and returns a float value representing the log-likelihood value for those parameters. The __call__ method should not be overwritten by subclasses unless you know what you are doing!

Parameters:

• self – Refer to the object itself

• parameter_values – Compute the covariance matrix

• interpolation_value_0 – Interpolate the covariance matrix along the first dimension

• interpolation_value_1 – Interpolate the covariance matrix

:param : Compute the covariance sum

Returns:

The log-likelihood function

class flip.likelihood.Prior(parameter_name=None)

parameter_name = None

class flip.likelihood.GaussianPrior(parameter_name=None, prior_mean=None, prior_standard_deviation=None)

Bases: Prior

prior_mean = None

prior_standard_deviation = None

__call__(parameter_values_dict)

class flip.likelihood.PositivePrior(parameter_name=None)

Bases: Prior

__call__(parameter_values_dict)

class flip.likelihood.UniformPrior(parameter_name=None, range=None)

Bases: Prior

range = None

__call__(parameter_values_dict)