def score_data(self, data, feature):
if len(data.domain.class_vars) != 1:
raise ValueError("ReliefF requires one single class")
if not data.domain.class_var.is_discrete:
raise ValueError("ReliefF supports classification; use RReliefF "
"for regression")
if len(data.domain.class_var.values
) == 1: # Single-class value non-problem
return 0 if feature else np.zeros(data.X.shape[1])
if isinstance(self.random_state, np.random.RandomState):
rstate = self.random_state
else:
rstate = np.random.RandomState(self.random_state)
from Orange.preprocess._relieff import relieff
weights = np.asarray(
relieff(
data.X,
data.Y,
self.n_iterations,
self.k_nearest,
np.array([a.is_discrete for a in data.domain.attributes]),
rstate,
))
if feature:
return weights[0]
return weights
def score_data(self, data, feature):
if len(data.domain.class_vars) != 1:
raise ValueError('ReliefF requires one single class')
if not data.domain.class_var.is_discrete:
raise ValueError('ReliefF supports classification; use RReliefF '
'for regression')
if len(data.domain.class_var.values) == 1: # Single-class value non-problem
return 0 if feature else np.zeros(data.X.shape[1])
from Orange.preprocess._relieff import relieff
weights = np.asarray(relieff(data.X, data.Y,
self.n_iterations, self.k_nearest,
np.array([a.is_discrete for a in data.domain.attributes])))
if feature:
return weights[0]
return weights
