regression trees up and rdy for testing

ropensci · Nov 13, 2023 · bdfa0cc · bdfa0cc
1 parent 8ccb638
commit bdfa0cc
Show file tree

Hide file tree

Showing 37 changed files with 2,245 additions and 884 deletions.
diff --git a/R/RcppExports.R b/R/RcppExports.R
@@ -37,8 +37,8 @@ compute_pred_prob_exported <- function(y, w) {
     .Call(`_aorsf_compute_pred_prob_exported`, y, w)
 }
 
-expand_y_clsf <- function(y, n_class) {
-    .Call(`_aorsf_expand_y_clsf`, y, n_class)
+compute_var_reduction_exported <- function(y_node, w_node, g_node) {
+    .Call(`_aorsf_compute_var_reduction_exported`, y_node, w_node, g_node)
 }
 
 is_col_splittable_exported <- function(x, y, r, j) {
@@ -73,11 +73,15 @@ cph_scale <- function(x, w) {
     .Call(`_aorsf_cph_scale`, x, w)
 }
 
-orsf_cpp <- function(x, y, w, tree_type_R, tree_seeds, loaded_forest, lincomb_R_function, oobag_R_function, n_tree, mtry, sample_with_replacement, sample_fraction, vi_type_R, vi_max_pvalue, leaf_min_events, leaf_min_obs, split_rule_R, split_min_events, split_min_obs, split_min_stat, split_max_cuts, split_max_retry, lincomb_type_R, lincomb_eps, lincomb_iter_max, lincomb_scale, lincomb_alpha, lincomb_df_target, lincomb_ties_method, pred_mode, pred_type_R, pred_horizon, pred_aggregate, oobag, oobag_eval_type_R, oobag_eval_every, pd_type_R, pd_x_vals, pd_x_cols, pd_probs, n_thread, write_forest, run_forest, verbosity) {
-    .Call(`_aorsf_orsf_cpp`, x, y, w, tree_type_R, tree_seeds, loaded_forest, lincomb_R_function, oobag_R_function, n_tree, mtry, sample_with_replacement, sample_fraction, vi_type_R, vi_max_pvalue, leaf_min_events, leaf_min_obs, split_rule_R, split_min_events, split_min_obs, split_min_stat, split_max_cuts, split_max_retry, lincomb_type_R, lincomb_eps, lincomb_iter_max, lincomb_scale, lincomb_alpha, lincomb_df_target, lincomb_ties_method, pred_mode, pred_type_R, pred_horizon, pred_aggregate, oobag, oobag_eval_type_R, oobag_eval_every, pd_type_R, pd_x_vals, pd_x_cols, pd_probs, n_thread, write_forest, run_forest, verbosity)
+expand_y_clsf <- function(y, n_class) {
+    .Call(`_aorsf_expand_y_clsf`, y, n_class)
+}
+
+compute_mse_exported <- function(y, w, p) {
+    .Call(`_aorsf_compute_mse_exported`, y, w, p)
 }
 
-compute_var_reduction <- function(y_node, w_node, g_node) {
-    .Call(`_aorsf_compute_var_reduction`, y_node, w_node, g_node)
+orsf_cpp <- function(x, y, w, tree_type_R, tree_seeds, loaded_forest, lincomb_R_function, oobag_R_function, n_tree, mtry, sample_with_replacement, sample_fraction, vi_type_R, vi_max_pvalue, leaf_min_events, leaf_min_obs, split_rule_R, split_min_events, split_min_obs, split_min_stat, split_max_cuts, split_max_retry, lincomb_type_R, lincomb_eps, lincomb_iter_max, lincomb_scale, lincomb_alpha, lincomb_df_target, lincomb_ties_method, pred_mode, pred_type_R, pred_horizon, pred_aggregate, oobag, oobag_eval_type_R, oobag_eval_every, pd_type_R, pd_x_vals, pd_x_cols, pd_probs, n_thread, write_forest, run_forest, verbosity) {
+    .Call(`_aorsf_orsf_cpp`, x, y, w, tree_type_R, tree_seeds, loaded_forest, lincomb_R_function, oobag_R_function, n_tree, mtry, sample_with_replacement, sample_fraction, vi_type_R, vi_max_pvalue, leaf_min_events, leaf_min_obs, split_rule_R, split_min_events, split_min_obs, split_min_stat, split_max_cuts, split_max_retry, lincomb_type_R, lincomb_eps, lincomb_iter_max, lincomb_scale, lincomb_alpha, lincomb_df_target, lincomb_ties_method, pred_mode, pred_type_R, pred_horizon, pred_aggregate, oobag, oobag_eval_type_R, oobag_eval_every, pd_type_R, pd_x_vals, pd_x_cols, pd_probs, n_thread, write_forest, run_forest, verbosity)
 }
 
diff --git a/R/orsf.R b/R/orsf.R
@@ -371,6 +371,7 @@ orsf <- function(data,
               pred_type = oobag_pred_type,
               oobag_pred_horizon = oobag_pred_horizon,
               oobag_eval_every = oobag_eval_every,
+              oobag_fun = oobag_fun,
               importance_type = importance,
               importance_max_pvalue = importance_max_pvalue,
               importance_group_factors = group_factors,
@@ -391,7 +392,7 @@ orsf <- function(data,
   tree_type,
   'survival' = do.call(ObliqueForestSurvival$new, args = args),
   'classification' = do.call(ObliqueForestClassification$new, args = args),
-  'regression' = stop("not ready yet")
+  'regression' = do.call(ObliqueForestRegression$new, args = args)
  )
 
  if(no_fit) return(object)