Figure_4.tif (616.37 kB)
AIC Model Comparison.
figure
posted on 2014-03-03, 03:29 authored by Yuan-Chi Tseng, Joshua I. Glaser, Eamon Caddigan, Alejandro LlerasAkaike information criterion (AIC) values are compared between several sub-models of the RDM. Each sub-model has one or two constrained parameters compared with the full model. This comparison is done for every subject, and the sum of AIC values is used for an across-subjects comparison. Note that AIC biases model selection towards models with more parameters.