statsmodels.discrete.discrete_model.CountResults.bic

method

CountResults.bic()

Bayesian information criterion. -2*llf + ln(nobs)*p where p is the number of regressors including the intercept.