def test_homer_shallowly():
'''a very shallow test for HOMER'''
n_classes = 4
X, y = make_ml_clf(n_samples=50, n_features=20,
n_classes=n_classes,
n_labels=2,
length=10, allow_unlabeled=False,
return_indicator=True,
random_state=123456)
model = HOMER(base_clf=OneVsRestClassifier(LinearSVC(random_state=0)),
k=3,
max_iter=20,
random_state=123456)
model.fit(X, y)
assert_equal(model._label_n, n_classes)
assert_equal(type(model._estimator_hierarchy[0]),
OneVsRestClassifier)
for clf in model._estimator_hierarchy[1]:
assert_equal(type(clf[0]), OneVsRestClassifier)
assert_equal(len(clf[1]), 0) # no more children clf
model.predict(X)
assert_equal(model._meta_y_hier[0].shape, (50, 2))
for y_hier in model._meta_y_hier[1]:
assert_equal(y_hier.shape, (50, 2))
def plot_2d(ax, n_labels=1, n_classes=3, length=50):
X, Y, p_c, p_w_c = make_ml_clf(
n_samples=150,
n_features=2,
n_classes=n_classes,
n_labels=n_labels,
length=length,
allow_unlabeled=False,
return_distributions=True,
random_state=RANDOM_SEED,
)
ax.scatter(X[:, 0],
X[:, 1],
color=COLORS.take((Y * [1, 2, 4]).sum(axis=1)),
marker=".")
ax.scatter(
p_w_c[0] * length,
p_w_c[1] * length,
marker="*",
linewidth=0.5,
edgecolor="black",
s=20 + 1500 * p_c**2,
color=COLORS.take([1, 2, 4]),
)
ax.set_xlabel("Feature 0 count")
return p_c, p_w_c
def plot_2d(ax, n_labels=1, n_classes=3, length=50):
X, Y, p_c, p_w_c = make_ml_clf(n_samples=150, n_features=2,
n_classes=n_classes, n_labels=n_labels,
length=length, allow_unlabeled=False,
return_distributions=True,
random_state=RANDOM_SEED)
ax.scatter(X[:, 0], X[:, 1], color=COLORS.take((Y * [1, 2, 4]
).sum(axis=1)),
marker='.')
ax.scatter(p_w_c[0] * length, p_w_c[1] * length,
marker='*', linewidth=.5, edgecolor='black',
s=20 + 1500 * p_c ** 2,
color=COLORS.take([1, 2, 4]))
ax.set_xlabel('Feature 0 count')
return p_c, p_w_c
from imblearn.metrics import classification_report_imbalanced
from sklearn.datasets import make_multilabel_classification as make_ml_clf
from sklearn.datasets import make_classification
from sklearn.svm import LinearSVC
from imblearn.ensemble import EasyEnsemble
from sklearn.cross_validation import train_test_split
from balance_data import cluster_centroids, edit_nearest_neribours, near_miss
from balance_data import adasyn, smote
from balance_data import smoteenn
RANDOM_STATE = 42
RANDOM_SEED = np.random.randint(2**10)
# Generate a dataset
X, y = make_ml_clf(n_classes=5,n_samples=5000, n_features=20,\
n_labels=1 ,allow_unlabeled=True,\
return_distributions=False,\
random_state=RANDOM_SEED)
"""
X, y = make_classification(n_classes=5, class_sep=5,\
weights=[0.1,0.3,0.1,0.05,0.45], n_informative=3, n_redundant=1, flip_y=0,\
n_features=20, n_clusters_per_class=5, n_samples=1000, random_state=10)
"""
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2)
y_train = [np.argmax(yt) for yt in y_train]
y_test = [np.argmax(yt) for yt in y_test]
print('Original train dataset shape {}'.format(Counter(y_train)))
print('Original test dataset shape {}'.format(Counter(y_test)))
sample_methods = ['no_sample','easy_ensemble','cluster_centroids','edit_nearest_neribours',\