def test_regressor_evals_result(loop): # noqa
with cluster() as (s, [a, b]):
with Client(s["address"], loop=loop):
a = dxgb.XGBRegressor()
X2 = da.from_array(X, 5)
y2 = da.from_array(y, 5)
a.fit(X2, y2, eval_metric="rmse", eval_set=[(X, y)])
evals_result = a.evals_result()
b = xgb.XGBRegressor()
b.fit(X, y, eval_metric="rmse", eval_set=[(X, y)])
assert_eq(evals_result, b.evals_result())
def test_regressor(loop): # noqa
with cluster() as (s, [a, b]):
with Client(s['address'], loop=loop):
a = dxgb.XGBRegressor()
X2 = da.from_array(X, 5)
y2 = da.from_array(y, 5)
a.fit(X2, y2)
p1 = a.predict(X2)
b = xgb.XGBRegressor()
b.fit(X, y)
assert_eq(p1, b.predict(X))
def test_regressor(xgboost_loop): # noqa
with cluster() as (s, [a, b]):
with Client(s['address'], loop=xgboost_loop):
a = dxgb.XGBRegressor()
X2 = da.from_array(X, 5)
y2 = da.from_array(y, 5)
weight1 = da.from_array(weight, 5)
a.fit(X2, y2, sample_weight=weight1)
p1 = a.predict(X2)
b = xgb.XGBRegressor()
b.fit(X, y, sample_weight=weight)
np.testing.assert_array_almost_equal(a.feature_importances_,
b.feature_importances_)
assert_eq(p1, b.predict(X))
def main():
object = ps.preprocess()
X_train, X_test, y_train, y_test = object.cleaning()
param_grid = {
'objective': ['binary:logistic'],
'nround': [1000],
'max_depth': [8]
}
estimator = dxgb.XGBRegressor()
grid_search = GridSearchCV(estimator, param_grid, verbose=2, cv=2, n_jobs=-1)
client = Client(processes=False)
start_time = time.time()
with joblib.parallel_backend("dask"):
grid_search.fit(X_train, y_train)
end_time = time.time()
grid_search.predict(X_test)
print ("time difference in GridSearchCV second XGBRegressor is %d seconds " % end_time)
client.shutdown()
def test_regressor_with_early_stopping(loop): # noqa
with cluster() as (s, [a, b]):
with Client(s["address"], loop=loop):
a = dxgb.XGBRegressor()
X2 = da.from_array(X, 5)
y2 = da.from_array(y, 5)
a.fit(
X2,
y2,
early_stopping_rounds=4,
eval_metric="rmse",
eval_set=[(X, y)],
)
p1 = a.predict(X2)
b = xgb.XGBRegressor()
b.fit(X, y, early_stopping_rounds=4, eval_metric="rmse", eval_set=[(X, y)])
assert_eq(p1, b.predict(X))
assert_eq(a.best_score, b.best_score)
def test_validation_weights_xgbregressor(loop): # noqa
from sklearn.datasets import make_regression
from sklearn.metrics import mean_squared_error
# prepare training and test data
X, y = make_regression(n_samples=2000, random_state=42)
with cluster() as (s, [a, b]):
with Client(s["address"], loop=loop):
X_train, X_test = X[:1600], X[1600:]
y_train, y_test = y[:1600], y[1600:]
dX_train = da.from_array(X_train)
dy_train = da.from_array(y_train)
dX_test = da.from_array(X_test)
reg = dxgb.XGBRegressor()
reg.fit(
dX_train,
dy_train, # sample_weight=weights_train,
)
preds = reg.predict(dX_test)
rng = np.random.RandomState(0)
weights_train = 100.0 + rng.rand(len(X_train))
weights_train = da.from_array(weights_train)
weights_test = 100.0 + rng.rand(len(X_test))
reg.fit(
dX_train,
dy_train,
sample_weight=weights_train,
sample_weight_eval_set=[weights_test],
)
preds2 = reg.predict(dX_test)
err = mean_squared_error(preds, y_test)
err2 = mean_squared_error(preds2, y_test)
assert err != err2
y_train = y_train.to_dask_array(lengths=True)
print("scaling")
#scaler = StandardScaler()
#scaler.fit(X_train)
#scaled_data = scaler.transform(X_train)
#X_test = scaler.transform(X_test)
print("training")
# In[ ]:
base_model = dxgb.XGBRegressor(objective='reg:squarederror',
tree_method='hist',
verbosity=3,
n_jobs=-1,
n_estimators=1000,
learning_rate=0.010,
max_depth=0,
max_leaves=4,
grow_policy='lossguide')
with joblib.parallel_backend('dask'):
base_model.fit(X_train, y_train.flatten())
#base_model.save_model('base_line_no_max_deph_lr_%f_%i.model'%(lr,leaves))
#
predictions = base_model.predict(X_test)
predictions = client.persist(predictions)
#
#print ("########")
#print ("R^2:",r2_score(y_test.compute(), predictions.compute()))
#print ("MAE:",mean_absolute_error(y_test.compute(), predictions.compute()))