Compare models using the Akaike information criterion.
AIC balances fit against complexity, penalising each additional parameter by 2. BIC penalises parameters more heavily as sample size grows, so it favours simpler models than AIC does. Only differences in AIC are meaningful; the absolute value depends on constants that cancel when comparing models on the same data.
AIC
AIC = 2k − 2 ln(L); BIC = k ln(n) − 2 ln(L); lower is better
AIC = 2k − 2 ln(L); BIC = k ln(n) − 2 ln(L); lower is better AIC balances fit against complexity, penalising each additional parameter by 2. BIC penalises parameters more heavily as sample size grows, so it favours simpler models than AIC does.
Only differences in AIC are meaningful; the absolute value depends on constants that cancel when comparing models on the same data.
This calculator takes 4 inputs: Log likelihood of the model, Number of parameters, Sample size, AIC of the competing model. The pre-filled defaults are a realistic starting point — replace them with figures from your own environment for a result you can act on.