update dir

__init__.py

@@ -1,2 +1,2 @@
-__all__=["ensemble"]
+__all__=["cgb"]
 

cgb/__init__.py

@@ -0,0 +1,4 @@
+from .cgb import cgb_clf
+from .cgb import cgb_reg
+
+__all__ = ["cgb_clf", "cgb_reg"]

test/__init__.py

@@ -0,0 +1 @@
+

test/tets_clf.py

@@ -1,2 +1,50 @@
-#%%
-from ensemble import cgb_clf
+from sklearn.model_selection import train_test_split
+from cgb import cgb_clf, cgb_reg
+import sklearn.datasets as dt
+import warnings
+
+warnings.simplefilter("ignore")
+
+
+def model(clf=True):
+
+ if clf:
+ X, y = dt.load_iris(return_X_y=True)
+
+ x_train, x_test, y_train, y_test = train_test_split(X,
+ y,
+ test_size=0.2)
+
+ model_ = cgb_clf(max_depth=5,
+ subsample=0.5,
+ max_features='sqrt',
+ learning_rate=0.05,
+ random_state=1,
+ criterion="mse",
+ loss="log_loss",
+ n_estimators=100)
+
+ else:
+ X, y = dt.make_regression(n_targets=3)
+
+ x_train, x_test, y_train, y_test = train_test_split(X,
+ y,
+ test_size=0.2)
+ model_ = cgb_reg(learning_rate=0.1,
+ subsample=1,
+ max_features="sqrt",
+ loss='ls',
+ n_estimators=100,
+ max_depth=3,
+ random_state=2)
+
+ model_.fit(x_train, y_train)
+ print(model_.score(x_test, y_test))
+
+
+if __name__ == "__main__":
+ print('clf')
+ model(clf=True)
+ print('-----')
+ print('reg')
+ model(clf=False)