statsmodels.tools.eval_measures.bic¶

statsmodels.tools.eval_measures.bic(llf, nobs, df_modelwc)[source]¶

Bayesian information criterion (BIC) or Schwarz criterion

Parameters:¶

llf{float, array_like}

value of the loglikelihood

nobsint

number of observations

df_modelwcint

number of parameters including constant

Returns:¶

bicfloat

information criterion

References

https://en.wikipedia.org/wiki/Bayesian_information_criterion