Open krzyzinskim opened 9 months ago
@krzyzinskim If I recall correctly, {randomForestSRC} even supports multi-output regressions (which is nice).
Hmm, in that case, I'll try to rethink the structure for handling non-single-number predictions so that multi-output regression can be also supported.
(Dito for XGBoost 2.0). The output structure can be like classification, I'd say: a list of matrices/arrays (?)
Adding support for survival models other than
ranger
(see #24). The source code for theranger
can be followed. Functionality forrandomForestSRC
should be relatively easy to implement.