“…For
‐SMEC model, we define the AIC and BIC following the proposal of Taavoni et al
38 as follows:
where
corresponds to the logarithm of the observed likelihood function
,
is the total number of parameters in the model, and
denotes the sample size. Note that other authors, such as Ibacache‐Pulgar et al
21 (see also Mattos et al
39 ), use the logarithm of the penalized likelihood function instead of the observed one. In our case, we compare both approaches, and the conclusions remain the same.…”