OCaml wrapper on top of a Random Forests implementation

UnixJunkie · September 19, 2018, 6:41am

Hello,

I made yet another wrapper.
It is here:

It can be used for classification and regression.
There is a test here:

UnixJunkie/oranger/blob/master/src/test.ml

open Printf

module L = BatList

module Score_label = struct
  type t = bool * float (* (label, pred_score) *)
  let get_label (l, _) = l
  let get_score (_, s) = s
end

module ROC = Cpm.MakeROC.Make(Score_label)


let main () =
  Log.(set_log_level DEBUG);
  Log.color_on ();
  let argc, args = CLI.init () in
  let verbose = CLI.get_set_bool ["-v"] args in
  let nb_trees = 500 in
  (* classification test *)

This file has been truncated. show original

This package relies on the C++ ranger implementation:

this package does not depend on R.

It is quite hackish and doesn’t support sparse data also (I need to know how ranger
supports them first).

Regards,
F.

Topic		Replies	Views
[ANN] random forests in OCaml via bindings to the R randomForest package Community	0	808	July 25, 2018
[ANN] Release of OCaml-R Community r , statistics	8	1250	November 22, 2018
A Random Forests classifier in OCaml Community machine-learning , regression , classification , random-forests	3	819	June 17, 2021
[ANN] first release of orf: OCaml Random Forests Community announce , machine-learning , regression , classification , random-forests	0	721	June 22, 2021
[ANN] New release of ocaml-r Community release , r , statistics	0	723	November 14, 2020

OCaml wrapper on top of a Random Forests implementation

Related Topics