LikelihoodCalculator

class remu.likelihood.LikelihoodCalculator(data_model, predictor, systematics='marginal')

Class that calculates the likelihoods of parameter sets.

Parameters

data_modelDataModel

Object that describes the statistical model of the data.

predictorPredictor

The object that translates parameter sets into reco expectation values.

systematics{‘marginal’, ‘profile’} or SystematicsConsumer, optional

Specifies how to handle systematic prediction uncertainties, i.e. multiple predictions from a single parameter set.

Notes

TODO

Methods

__call__(*args, **kwargs)	Call self as a function.
compose(predictor)	Return a new LikelihoodCalculator with the composed Predictor.
fix_parameters(fix_values)	Return a new LikelihoodCalculator with fewer free parameters.
generate_toy_likelihood_calculators(parameters)	Generate LikelihoodCalculator objects with randomly varied data.
log_likelihood(*args, **kwargs)	Calculate the log likelihood of the given parameters.

compose(predictor)

Return a new LikelihoodCalculator with the composed Predictor.

Parameters

predictorPredictor

The predictor of the calculator will be composed with this.

fix_parameters(fix_values)

Return a new LikelihoodCalculator with fewer free parameters.

Parameters

fix_values

generate_toy_likelihood_calculators(parameters, N=1, **kwargs)

Generate LikelihoodCalculator objects with randomly varied data.

Accepts only single set of parameters.

Returns

toy_calculatorslist of LikelihoodCalculator

log_likelihood(*args, **kwargs)

Calculate the log likelihood of the given parameters.

Passes all arguments to the predictor.