The training data are used for parameter estimation, the validation data are used for model selection (e.g. to select the optimal penalty strength in the penalized regression methods) and the testing data are used to estimate prediction accuracy.
Each modeling regime was fit independently with a resulting likelihood used for model selection involving Akaike information criterion (AIC).
Two criteria were used for model selection.
Additional odorant activity data we used for model selection.
In this research, BIC is used for model selection.
The data-driven method used for model selection in this paper found a reversible model for [18F]HX4.
Likelihood based procedures like Akaike Information criterion and Bayes Information Criterion are usually used for model selection in GLM.
As the sample had already been used for model selection, this estimate of model performance is overly optimistic.
As far as we are aware, nested cross-validation is the best non-parametric approach for model assessment when cross-validation is used for model selection.
As Stone [2] pointed out, cross-validation can be used for model selection and for model assessment, but the two tasks require different cross-validation approaches.
The performances of fit indices used for model selection in cross-sectional mixture modeling with nonnormally distributed indicators were examined in two studies using Monte Carlo methods.
