statsmodels.discrete.discrete_model.BinaryResults.bic¶ BinaryResults.bic¶ Bayesian information criterion. -2*llf + ln(nobs)*p where p is the number of regressors including the intercept.