I finished implementing a classifier and regressor using Random Forests
(seminal paper: https://link.springer.com/article/10.1023/A:1010933404324):
- this is somewhat slow; especially the classifier
(and I don’t know so much how to accelerate it; probably two orders of magnitude slower
- this is not super generic (int IntMap sparse features only; i.e. a sparse vector of integers
represents a sample).
The package is now available in opam (opam install orf).
Two interfaces are exposed:
RFC (for classification)
RFR (for regression)
The test file shows some usage examples:
If you want to help, I tried to flag a few things for the near future:
If you use it and if it is useful to you, I would be happy to know.