def test_boston_dataset():
boston = load_boston()
X_train, X_test, y_train, y_test = train_test_split(boston.data,
boston.target,
random_state=42)
mapper = BinMapper(random_state=42)
X_train_binned = mapper.fit_transform(X_train)
X_test_binned = mapper.transform(X_test)
gradients = y_train.astype(np.float32)
hessians = np.ones(1, dtype=np.float32)
min_samples_leaf = 8
max_leaf_nodes = 31
grower = TreeGrower(X_train_binned,
gradients,
hessians,
min_samples_leaf=min_samples_leaf,
max_leaf_nodes=max_leaf_nodes)
grower.grow()
predictor = grower.make_predictor(bin_thresholds=mapper.bin_thresholds_)
assert r2_score(y_train, predictor.predict_binned(X_train_binned)) > 0.75
assert r2_score(y_test, predictor.predict_binned(X_test_binned)) > 0.65
assert_allclose(predictor.predict(X_train),
predictor.predict_binned(X_train_binned))
assert_allclose(predictor.predict(X_test),
predictor.predict_binned(X_test_binned))
assert r2_score(y_train, predictor.predict(X_train)) > 0.75
assert r2_score(y_test, predictor.predict(X_test)) > 0.65
def test_boston_dataset(max_bins):
boston = load_boston()
X_train, X_test, y_train, y_test = train_test_split(
boston.data, boston.target, random_state=42)
mapper = BinMapper(max_bins=max_bins, random_state=42)
X_train_binned = mapper.fit_transform(X_train)
X_test_binned = mapper.transform(X_test)
# Init gradients and hessians to that of least squares loss
gradients = -y_train.astype(np.float32)
hessians = np.ones(1, dtype=np.float32)
min_samples_leaf = 8
max_leaf_nodes = 31
grower = TreeGrower(X_train_binned, gradients, hessians,
min_samples_leaf=min_samples_leaf,
max_leaf_nodes=max_leaf_nodes, max_bins=max_bins,
n_bins_per_feature=mapper.n_bins_per_feature_)
grower.grow()
predictor = grower.make_predictor(bin_thresholds=mapper.bin_thresholds_)
assert r2_score(y_train, predictor.predict_binned(X_train_binned)) > 0.85
assert r2_score(y_test, predictor.predict_binned(X_test_binned)) > 0.70
assert_allclose(predictor.predict(X_train),
predictor.predict_binned(X_train_binned))
assert_allclose(predictor.predict(X_test),
predictor.predict_binned(X_test_binned))
assert r2_score(y_train, predictor.predict(X_train)) > 0.85
assert r2_score(y_test, predictor.predict(X_test)) > 0.70
def test_bin_mapper_random_data(n_bins):
n_samples, n_features = DATA.shape
expected_count_per_bin = n_samples // n_bins
tol = int(0.05 * expected_count_per_bin)
mapper = BinMapper(max_bins=n_bins, random_state=42).fit(DATA)
binned = mapper.transform(DATA)
assert binned.shape == (n_samples, n_features)
assert binned.dtype == np.uint8
assert_array_equal(binned.min(axis=0), np.array([0, 0]))
assert_array_equal(binned.max(axis=0), np.array([n_bins - 1, n_bins - 1]))
assert len(mapper.numerical_thresholds_) == n_features
for i in range(len(mapper.numerical_thresholds_)):
assert mapper.numerical_thresholds_[i].shape == (n_bins - 1, )
assert mapper.numerical_thresholds_[i].dtype == DATA.dtype
assert np.all(mapper.n_bins_per_feature_ == n_bins)
# Check that the binned data is approximately balanced across bins.
for feature_idx in range(n_features):
for bin_idx in range(n_bins):
count = (binned[:, feature_idx] == bin_idx).sum()
assert abs(count - expected_count_per_bin) < tol