Genetics Selection Evolution

Table 1 Classification error rates using naïve Bayes (NB), Bayesian networks (BN) and neural networks (NN) in five categorization schemes (K = 2, 3, 4, 5 and 10), based on the final SNP subsets selected

From: Comparison of classification methods for detecting associations between SNPs and chick mortality

		Number of categories (K)
Stratum^a	Classifier	K = 2	K = 3	K = 4	K = 5	K = 10
EL	NB	0.124	0.225	0.314	0.329	0.523
	BN	0.207	0.437	0.649	0.674	0.813
	NN	0.270	0.295	0.543	0.662	0.813
EH	NB	0.116	0.212	0.330	0.397	0.506
	BN	0.228	0.422	0.653	0.688	0.820
	NN	0.185	0.364	0.560	0.623	0.827
LL	NB	0.132	0.221	0.375	0.408	0.523
	BN	0.225	0.403	0.545	0.709	0.824
	NN	0.221	0.401	0.588	0.610	0.831
LH	NB	0.151	0.252	0.338	0.405	0.494
	BN	0.261	0.438	0.532	0.681	0.816
	NN	0.278	0.381	0.530	0.534	0.793

^a EL = early age-low hygiene; EH = early age-high hygiene; LL = late age-low hygiene; LH = late age-high hygiene.

Back to article page

ISSN: 1297-9686

Contact us

Submission enquiries: helene.hayes@inrae.fr
General enquiries: info@biomedcentral.com