def test_robust_scale_axis1():
X = iris.data
X_trans = robust_scale(X, axis=1)
assert_array_almost_equal(np.median(X_trans, axis=1), 0)
q = np.percentile(X_trans, q=(25, 75), axis=1)
iqr = q[1] - q[0]
assert_array_almost_equal(iqr, 1)
def test_robust_scale_axis1():
X = iris.data
X_trans = robust_scale(X, axis=1)
assert_array_almost_equal(np.median(X_trans, axis=1), 0)
q = np.percentile(X_trans, q=(25, 75), axis=1)
iqr = q[1] - q[0]
assert_array_almost_equal(iqr, 1)
def decision_function(self, X):
""" Compute the (weighted) sum of votes.
Parameters
----------
X : np.ndarray, List
Data in the form of rows x columns = samples x features
Returns
-------
The (weighted) sum of votes.
"""
if self.centroid_weighting:
votes = robust_scale(abs(X - self.averages[0, :]) - abs(X - self.averages[1, :]),
with_centering=False, axis=1)
else:
votes = (abs(X - self.averages[0, :]) > abs(X - self.averages[1, :])) - 0.5
if self.stats_weighting is None:
dec = np.sum(votes, 1) / votes.shape[1]
else:
dec = _weighted_sum(votes, self.feature_importances_) / votes.shape[1]
return dec
