Characterization of microbiota signatures in Iberian pig strains using machine learning algorithms

Table 1 AUROC scores ± SD of the best model before and after applying feature selection, and the AUROC mean percentage increase per scenario, considering all models

Scenario	No FS		FS		Mean increase
Scenario	AUROC	Model	AUROC	Model	Mean increase
Four genetic groups	0.64 ± 0.04	CB	0.68 ± 0.04	RF	6.9%
Purebred	0.77 ± 0.07	CB	0.83 ± 0.06	SVM	6.5%
Maternal	0.66 ± 0.06	XGB	0.70 ± 0.06	CB	4.3%
Paternal	0.68 ± 0.05	CB	0.74 ± 0.05	CB	8.7%
Heterosis	0.58 ± 0.06	CB	0.66 ± 0.06	CB	11.6%

Abbreviations; FS: Feature selection, CB: Catboost classifier, XGB: XGboost classifier, RF: Random Forest classifier, SVM: Support vector machine classifier

ISSN: 2524-4671