BestModelBIC

BestModelBIC(*args)

Select the best model according to the Bayesian information criterion.

Parameters:
sample2-d sequence of float

Tested sample.

modelslist of Distribution or DistributionFactory

Tested distributions.

Returns:
best_modelDistribution

The distribution that fits the sample best according to the criterion. This may raise a warning if the best model does not perform well.

best_bicfloat

The Bayesian information criterion with the best model.

Notes

The best model is the one which achieves the smallest BIC value. In case of ties, the order in the list matters: the first one which minimizes the BIC in the list is selected.

Examples

>>> import openturns as ot
>>> ot.RandomGenerator.SetSeed(0)
>>> distribution = ot.Normal()
>>> sample = distribution.getSample(30)
>>> tested_distributions = [ot.ExponentialFactory(), ot.NormalFactory()]
>>> best_model, best_bic = ot.FittingTest.BestModelBIC(sample, tested_distributions)
>>> print(best_model)
Normal(mu = -0.0944924, sigma = 0.989808)

Examples using the function

Estimate a multivariate distribution

Estimate a multivariate distribution

Select fitted distributions

Select fitted distributions