frambi_test
Test and compare multiple model variants

Usage:

[best_idx, BIC_diff] = frambi_test(metrics);
[best_idx, BIC_diff] = frambi_test(metrics, names);
[best_idx, BIC_diff, best_name] = frambi_test(metrics, names);
[...] = frambi_test(metrics, names, do_disp);

Input parameters:

metrics

Cell array with each cell containing a structure with the statistical results of the model variants as obtained from frambi_estimate:

loglik: Log-likelihood of the estimation.
BIC: Bayesian information criterion of the estimation.

names

Optional cell array containing names of each model variant. Default: Index of the variant in metrics.

do_disp

Optional binary flag to display the information about the winning variant and its difference in BIC to the second best variant. Default: True.

Output parameters:

`best_idx`	Index of the model variant with the lowest BIC.
`BIC_diff`	Vector with BIC differences between the best variant and all variants (including the best).
`best_name`	Name of the model with the lowest BIC value.

Description:

frambi_test(..) identifies the model variant with the lowest summed BIC, and optionally displays a statement about the winning variant. It returns the index and name of the best variant along with the BIC differences from the best variant to all variants (including the best variant).

References:

R. Barumerli and P. Majdak. FrAMBI: A Software Framework for Auditory Modeling Based on Bayesian Inference. under review at Neuroinformatics, 2024.