do8572 / MLDS

MLDS course code.
0 stars 0 forks source link

Claim 2.3 hsRF much faster to fit than BART (10 - 15 times faster) #10

Open do8572 opened 1 year ago

do8572 commented 1 year ago

2.4 hsRF much faster to fit than BART (10 - 15 times faster)

We also compare hsRF to the predictive performance of Bayesian Additive Regression Trees (BART) (Chipman et al., 2010), and observe that hsRF and BART are compa-rable in terms of prediction performance. However, hsRF is much faster to fit than BART (typically 10-15 times faster) and we can also apply HS to BART (see Appendix S4.5).