def train_cuml_classifier(data, labels, depth=25, trees=100):
"""Train SKLearn classification model"""
model = cuRFC(
max_depth=depth, n_estimators=trees, random_state=0
)
return model.fit(data, labels)
def _construct_rf(
n_estimators,
random_state,
**kwargs
):
return cuRFC(
n_estimators=n_estimators,
random_state=random_state,
**kwargs
)
def RandomForest_train(self, n_estimators):
# training AdaBoost for each scored label
self.y_pred = np.zeros(
(self.X_val.shape[0], len(self.targets_scored_col_name)))
for i in tqdm(range(len(self.targets_scored_col_name))):
this_target_col_name = self.targets_scored_col_name[i]
# if samples < 5 in this label, pass
if self.y_train[this_target_col_name].values.sum() < 5:
self.y_pred[:, i] = np.zeros(len(self.X_val))
else:
self.model = cuRFC(
n_estimators=n_estimators, # Number of trees in the forest
max_depth=32, # Maximum tree depth
max_features=
'auto', # Ratio of number of features (columns) to consider per node split
)
self.model.fit(self.X_train[self.features],
self.y_train[this_target_col_name])
self.y_pred[:, i] = cupy.asnumpy(
self.model.predict_proba(
self.X_val[self.features]).values)[:, 1]
def _func_build_rf(
n,
n_estimators,
max_depth,
handle,
max_features,
n_bins,
split_algo,
split_criterion,
min_rows_per_node,
bootstrap,
bootstrap_features,
type_model,
verbose,
rows_sample,
max_leaves,
n_streams,
quantile_per_tree,
dtype,
r,
):
return cuRFC(
n_estimators=n_estimators,
max_depth=max_depth,
handle=handle,
max_features=max_features,
n_bins=n_bins,
split_algo=split_algo,
split_criterion=split_criterion,
min_rows_per_node=min_rows_per_node,
bootstrap=bootstrap,
bootstrap_features=bootstrap_features,
type_model=type_model,
verbose=verbose,
rows_sample=rows_sample,
max_leaves=max_leaves,
n_streams=n_streams,
quantile_per_tree=quantile_per_tree,
gdf_datatype=dtype,
)
def _construct_rf(n_estimators, seed, **kwargs):
return cuRFC(n_estimators=n_estimators, seed=seed, **kwargs)
