Table 3.

Prediction accuracies of different models for all three growth traits.

Model *	BW	WW	YW
GBLUP	0.447 ± 0.023	0.373 ± 0.019	0.321 ± 0.022
Bayes A	0.426 ± 0.018	0.375 ± 0.012	0.303 ± 0.013
Bayes B	0.21 ± 0.010	0.26 ± 0.022	0.22 ± 0.015
RF_default	0.350 ± 0.025	0.405 ± 0.003	0.261 ± 0.001
SVM_default	0.334 ± 0.014	0.360 ± 0.021	0.280 ± 0.027
RF_optimized	0.421 ± 0.039	0.451 ± 0.011	0.306 ± 0.028
SVM_optimized	0.406 ± 0.034	0.380 ± 0.018	0.311 ± 0.004
MLP	0.408 ± 0.021	0.402 ± 0.027	0.304 ± 0.014
CNN	0.433 ± 0.034	0.419 ± 0.024	0.303 ± 0.011

* Prediction accuracy of the model is calculated as Pearson’s correlation between corrected phenotypes and direct genomic values in the validation dataset.