mesmer.distrib.ConditionalDistribution

mesmer.distrib.ConditionalDistribution#

class mesmer.distrib.ConditionalDistribution(expression, *, minimize_options=None, second_minimizer=None, threshold_min_proba=1e-09)#
__init__(expression, *, minimize_options=None, second_minimizer=None, threshold_min_proba=1e-09)#

A conditional distribution.

Parameters:
  • Expression (class py:class:Expression) – Expression defining the conditional distribution.

  • minimize_options (py:class:`MinimizeOptions | None, default: MinimizeOptions()) – Class defining the optimizer options used during first guess and training of distributions. If not passed uses “Nelder-Mead” minimizer with default settings.

  • second_minimizer (class py:class:MinimizeOptions | None, default: None) – Run a second minimization algorithm for all steps. method="Powell" is recommended. It can be beneficial to run more than one minimization to get a more stable estimate.

Methods

__init__(expression, *[, minimize_options, ...])

A conditional distribution.

compute_quality_scores(predictors, target, ...)

Compute scores for fit coefficients.

find_first_guess(predictors, target, weights)

Find a first guess for all coefficients of a conditional distribution

fit(predictors, target, weights, first_guess, *)

fit conditional distribution over all gridpoints.

from_dataset(ds)

set coefficients from a dataset with default solver options

from_netcdf(filename, **kwargs)

read coefficients from a netCDF file with default solver options

to_netcdf(filename, **kwargs)

save coefficients dataset to a netCDF file

Attributes

coefficients

The coefficients of this conditional distribution.