2.2 hsRF achieves maximum performance with fewer trees than RF (5 times fewer trees)
Moreover, hsRF tends to achieve its maximum performance with fewer trees than RF without regularization; as a consequence, RF with HS is of-ten able to achieve the same performance with an ensemble that is five times smaller, allowing us to achieve large sav-ings in computational resources.
2.2 hsRF achieves maximum performance with fewer trees than RF (5 times fewer trees)