A comparison of random forests, boosting and support vector machines for genomic selection

Accuracy was highest for boosting, intermediate for SVMs and lowest for RF but differed little among the three methods and relative to ridge regression BLUP (RR-BLUP).