def test_predict_sparse(self):
sparse_data = self.housing.to_sparse()
booster = CatGBRegressor()
model = booster(sparse_data)
pred = model(sparse_data)
self.assertEqual(pred.shape, (len(sparse_data),))
self.assertGreater(all(pred), 0)
def test_set_params(self):
booster = CatGBRegressor(n_estimators=42, max_depth=4)
self.assertEqual(booster.params["n_estimators"], 42)
self.assertEqual(booster.params["max_depth"], 4)
model = booster(self.housing)
params = model.cat_model.get_params()
self.assertEqual(params["n_estimators"], 42)
self.assertEqual(params["max_depth"], 4)
def test_default_parameters_reg(self):
data = Table("housing")
booster = CatGBRegressor()
model = booster(data)
params = model.cat_model.get_all_params()
self.assertEqual(self.editor.n_estimators, 100)
self.assertEqual(params["iterations"], 1000)
self.assertEqual(params["depth"], self.editor.max_depth)
self.assertEqual(params["l2_leaf_reg"], self.editor.lambda_)
self.assertEqual(params["rsm"], self.editor.colsample_bylevel)
self.assertEqual(self.editor.learning_rate, 0.3)
def test_scorer(self):
booster = CatGBRegressor()
self.assertIsInstance(booster, Scorer)
booster.score(self.housing)
def test_predict_numpy(self):
booster = CatGBRegressor()
model = booster(self.housing)
pred = model(self.housing.X)
self.assertEqual(pred.shape, (len(self.housing),))
self.assertGreater(all(pred), 0)
def test_predict_single_instance(self):
booster = CatGBRegressor()
model = booster(self.housing)
for ins in self.housing:
pred = model(ins)
self.assertGreater(pred, 0)
def test_GBTrees(self):
booster = CatGBRegressor()
cv = CrossValidation(k=10)
results = cv(self.housing, [booster])
RMSE(results)