diff --git a/multimodal/boosting/combo.py b/multimodal/boosting/combo.py
index bad1f32ce2c786bc7a95b6631ce6f3b16af0ac9d..5f4ef5346c3b6946acf4cfcda3617fbf09884953 100644
--- a/multimodal/boosting/combo.py
+++ b/multimodal/boosting/combo.py
@@ -193,7 +193,7 @@ class MuComboClassifier(BaseEnsemble, ClassifierMixin, UBoosting):
# n_yi = np.unique(y, return_inverse=True)
cost = np.ones((n_views, n_samples, n_classes))
score_function = np.zeros((n_views, n_samples, n_classes))
- n_yi_s = np.zeros(n_classes, dtype=np.int)
+ n_yi_s = np.zeros(n_classes, dtype=int)
for indice_class in range(n_classes):
# n_yi number of examples of the class y_i
n_yi = np.where(y==indice_class)[0].shape[0]
@@ -224,14 +224,14 @@ class MuComboClassifier(BaseEnsemble, ClassifierMixin, UBoosting):
def _indicatrice(self, predicted_classes, y_i):
n_samples = y_i.shape[0]
- indicate_ones = np.zeros((self.n_views_, n_samples, self.n_classes_), dtype=np.int)
- indicatrice_one_yi = np.zeros((self.n_views_, n_samples, self.n_classes_), dtype=np.int)
+ indicate_ones = np.zeros((self.n_views_, n_samples, self.n_classes_), dtype=int)
+ indicatrice_one_yi = np.zeros((self.n_views_, n_samples, self.n_classes_), dtype=int)
indicate_ones[np.arange(self.n_views_)[:, np.newaxis],
np.arange(n_samples)[np.newaxis, :],
predicted_classes[np.arange(self.n_views_), :]] = 1
indicate_ones[:, np.arange(n_samples), y_i] = 0
indicatrice_one_yi[:, np.arange(n_samples), y_i] = 1
- delta = np.ones((self.n_views_, n_samples, self.n_classes_), dtype=np.int)
+ delta = np.ones((self.n_views_, n_samples, self.n_classes_), dtype=int)
delta[:, np.arange(n_samples), y_i] = -1
return indicate_ones, indicatrice_one_yi, delta
@@ -451,12 +451,12 @@ class MuComboClassifier(BaseEnsemble, ClassifierMixin, UBoosting):
# there is only one single class in the data).
self.estimators_ = []
self.estimator_weights_alpha_ = np.array([], dtype=np.float64)
- self.estimator_weights_beta_ = np.zeros((self.n_iterations_, n_views), dtype=np.float)
+ self.estimator_weights_beta_ = np.zeros((self.n_iterations_, n_views), dtype=float)
self.estimator_errors_ = np.array([], dtype=np.float64)
return
self.estimators_ = []
self.estimator_weights_alpha_ = np.zeros((self.n_iterations_, n_views), dtype=np.float64)
- self.estimator_weights_beta_ = np.zeros((self.n_iterations_, n_views, self.n_classes_), dtype=np.float)
+ self.estimator_weights_beta_ = np.zeros((self.n_iterations_, n_views, self.n_classes_), dtype=float)
self.estimator_errors_ = np.zeros((n_views, self.n_iterations_), dtype=np.float64)
random_state = check_random_state(self.random_state)
diff --git a/multimodal/datasets/data_sample.py b/multimodal/datasets/data_sample.py
index b65fe0ee134b17a5e00709bf6bce2bbe23c55f3b..c44b360e9c2d808f51bcfa22acfd32e80b1db8fa 100644
--- a/multimodal/datasets/data_sample.py
+++ b/multimodal/datasets/data_sample.py
@@ -83,6 +83,7 @@ import numpy as np
import numpy.ma as ma
import scipy.sparse as sp
+
class MultiModalData(metaclass=ABCMeta):
@staticmethod
@@ -99,7 +100,7 @@ class MultiModalData(metaclass=ABCMeta):
n_views = views_ind.shape[0]-1
else:
if views_ind.ndim == 1:
- if not views_ind.dtype == np.object:
+ if not views_ind.dtype == object:
raise ValueError("The format of views_ind is not "
+ "supported.")
for ind, val in enumerate(views_ind):
@@ -143,7 +144,7 @@ class MultiModalData(metaclass=ABCMeta):
n_views = views_ind.shape[0]-1
else:
if views_ind.ndim == 1:
- if not views_ind.dtype == np.object:
+ if not views_ind.dtype == object:
raise ValueError("The format of views_ind is not "
+ "supported.")
for ind, val in enumerate(views_ind):
@@ -428,7 +429,7 @@ class MultiModalArray(np.ndarray, MultiModalData):
hasattr(new_data, "shape") and len(new_data.shape) > 0:
obj = np.asarray(new_data.data).view(cls)
else:
- obj = np.recarray.__new__(cls, shape=(0, 0), dtype=np.float)
+ obj = np.recarray.__new__(cls, shape=(0, 0), dtype=float)
obj.view_mode_ = view_mode
obj.views_ind = views_ind
obj.shapes_int = shapes_int
@@ -439,7 +440,7 @@ class MultiModalArray(np.ndarray, MultiModalData):
def _for_data(cls, data):
n_views = data.shape[0]
index = 0
- views_ind = np.empty(n_views + 1, dtype=np.int)
+ views_ind = np.empty(n_views + 1, dtype=int)
views_ind[0] = 0
shapes_int = []
new_data = np.ndarray([])
diff --git a/multimodal/tests/test_combo.py b/multimodal/tests/test_combo.py
index f8d085dfa2138313b8060fc51eb81456dc87876f..5b864e62c9e581cf0d638591b8c21dd537d6f22e 100644
--- a/multimodal/tests/test_combo.py
+++ b/multimodal/tests/test_combo.py
@@ -268,7 +268,7 @@ class TestMuComboClassifier(unittest.TestCase):
cost_Tminus1_vue, A, b, G, h, l))
self.assertEqual(solver.shape, (9,1))
s_r = np.sum(solver.reshape(3,3), axis=1)
- np.testing.assert_almost_equal(s_r, np.ones(3, dtype=np.float), 9)
+ np.testing.assert_almost_equal(s_r, np.ones(3, dtype=float), 9)
def test_solver_compute_betas(self):
clf = MuComboClassifier()
@@ -289,13 +289,13 @@ class TestMuComboClassifier(unittest.TestCase):
[ 5.45938775e+01, 4.03397223e+02, -4.57991101e+02],
[ 1.48401545e+02, -1.48426438e+02, 2.48935342e-02],
[ 1.09663316e+03, 2.71828184e+00, -1.09935144e+03]]])
- score_function_Tminus1 =10 *np.ones((3,4,3), dtype =np.float)
+ score_function_Tminus1 =10 *np.ones((3,4,3), dtype =float)
alphas = np.array([0.5, 1., 2.])
predicted_classes = np.array([[0, 0, 1, 1], [0, 1, 0, 2], [2, 2, 0, 0]])
y = np.array([0, 1, 2, 0])
betas = clf._compute_betas(alphas, y, score_function_Tminus1, predicted_classes)
self.assertEqual(betas.shape, (3,3))
- np.testing.assert_almost_equal(np.sum(betas, axis =1), np.ones(3, dtype=np.float), 9)
+ np.testing.assert_almost_equal(np.sum(betas, axis =1), np.ones(3, dtype=float), 9)
self.assertTrue(np.all(betas <= 1) )
self.assertTrue(np.all(betas >= 0) )
@@ -336,7 +336,7 @@ class TestMuComboClassifier(unittest.TestCase):
[[-2, 1, 0.5], [1, 1, -1], [1, -2, 0.5], [1, 1, -1]],
[[-2, 1, 0.5], [1, 1, -1], [1, -2, 0.5], [1, 1, -1]]],
dtype=np.float64)
- score_function_Tminus1 =10 *np.ones((3,4,3), dtype=np.float)
+ score_function_Tminus1 =10 *np.ones((3,4,3), dtype=float)
clf = MuComboClassifier()
clf.n_views_ = 3
clf.n_classes_ = 3
diff --git a/multimodal/tests/test_mumbo.py b/multimodal/tests/test_mumbo.py
index af9c5c69f222c26ad892c1cbf2e4bfb68104574b..fd626247b536fb2013e1160ac2dfec55a2b584e8 100644
--- a/multimodal/tests/test_mumbo.py
+++ b/multimodal/tests/test_mumbo.py
@@ -74,7 +74,7 @@ class TestMumboClassifier(unittest.TestCase):
X[X < .8] = 0
X_csr = csr_matrix(X)
clf = MumboClassifier()
- y = (4 * rng.rand(40)).astype(np.int)
+ y = (4 * rng.rand(40)).astype(int)
clf.fit(X_csr, y)
def test_init_var(self):
@@ -450,15 +450,15 @@ class TestMumboClassifier(unittest.TestCase):
clf = MumboClassifier()
self.assertRaises(ValueError, clf.fit, X, y, views_ind)
- views_ind = np.array([[-1], [1, 2]])
+ views_ind = np.array([np.array([-1]), np.array([1, 2])])
clf = MumboClassifier()
self.assertRaises(ValueError, clf.fit, X, y, views_ind)
- views_ind = np.array([[3], [1, 2]])
+ views_ind = np.array([np.array([3]), np.array([1, 2])])
clf = MumboClassifier()
self.assertRaises(ValueError, clf.fit, X, y, views_ind)
- views_ind = np.array([[0.5], [1, 2]])
+ views_ind = np.array([np.array([0.5]), np.array([1, 2])])
clf = MumboClassifier()
self.assertRaises(ValueError, clf.fit, X, y, views_ind)
@@ -520,7 +520,7 @@ class TestMumboClassifier(unittest.TestCase):
X = np.array([[0., 0.5, 0.7], [1., 1.5, 1.7], [2., 2.5, 2.7]])
y = np.array([1, 1, 1])
- views_ind = np.array([[0, 2], [1]])
+ views_ind = np.array([np.array([0, 2]), np.array([1])])
clf = MumboClassifier()
clf.fit(X, y, views_ind)
np.testing.assert_equal(clf.predict(X), y)
@@ -573,7 +573,7 @@ class TestMumboClassifier(unittest.TestCase):
np.array([0, 1]))
self.assertEqual(clf.decision_function(X).shape, y.shape)
- views_ind = np.array([[2, 0], [1]])
+ views_ind = np.array([np.array([2, 0]), np.array([1])])
clf = MumboClassifier()
clf.fit(X, y, views_ind)
np.testing.assert_equal(clf.predict(X), y)
@@ -598,7 +598,7 @@ class TestMumboClassifier(unittest.TestCase):
np.testing.assert_equal(clf.predict(data), np.array([0, 1]))
self.assertEqual(clf.decision_function(X).shape, y.shape)
- views_ind = np.array([[2, 0], [1], [3, 4]])
+ views_ind = np.array([np.array([2, 0]), np.array([1]), np.array([3, 4])])
clf = MumboClassifier()
clf.fit(X, y, views_ind)
np.testing.assert_equal(clf.predict(X), y)
@@ -626,7 +626,7 @@ class TestMumboClassifier(unittest.TestCase):
np.testing.assert_equal(clf.predict(data), np.array([0, 1, 2]))
self.assertEqual(clf.decision_function(X).shape, (X.shape[0], 3))
- views_ind = np.array([[1, 0], [2], [3, 4]])
+ views_ind = np.array([np.array([1, 0]), np.array([2]), np.array([3, 4])])
clf = MumboClassifier()
clf.fit(X, y, views_ind)
np.testing.assert_equal(clf.predict(X), y)
diff --git a/multimodal/tests/test_mvml.py b/multimodal/tests/test_mvml.py
index 83c6ea3125032a0574398185fbe3b220da427453..3a27d97ebdbe3114b12e306d5d04777af99aa7fc 100644
--- a/multimodal/tests/test_mvml.py
+++ b/multimodal/tests/test_mvml.py
@@ -270,7 +270,7 @@ class MVMLTest(unittest.TestCase):
def test_classifier(self):
pass
- # return check_estimator(MVML)
+ # return check_estimator(MVML())
def test_check_kernel(self):
clf = MVML()
diff --git a/setup.py b/setup.py
index 602cb5d6950ff9949c8d3c8ac32ceac2bcea7b01..d9f23f30cf933a95dfba03dd8c9e40aa91de1b77 100644
--- a/setup.py
+++ b/setup.py
@@ -176,7 +176,7 @@ def setup_package():
keywords = ('machine learning, supervised learning, classification, '
'ensemble methods, boosting, kernel')
packages = find_packages(exclude=['*.tests'])
- install_requires = ['scikit-learn==0.24.2', 'numpy', 'scipy', 'cvxopt' ]
+ install_requires = ['scikit-learn>=0.24.2', 'numpy', 'scipy', 'cvxopt' ]
python_requires = '>=3.6'
extras_require = {
'dev': ['pytest', 'pytest-cov'],