Skip to main content
. 2017 Dec 8;4(6):ENEURO.0311-16.2017. doi: 10.1523/ENEURO.0311-16.2017

Table 5.

Comparison of models for corpus and relative frequency

df AIC BIC logLik
m.rel.index 8 2043373 2043457 −1021678
m.freq.index 8 2043326 2043410 −1021655

Both models yield similar fits as evidenced by log-likelihood, AIC, and BIC. Model names reflect the predictor used; ‘rel’ refers to relative frequency and ‘freq’ refers to corpus frequency.