def _check_is_fitted(self):
"""Check if the classifier is trained (fitted).
Raises
------
NotFittedError
If the classifier is not fitted.
"""
check_is_fitted(self, ['classes', 'n_features'])
def _check_is_fitted(self):
"""Check if the preprocessor is trained (fitted).
Raises
------
NotFittedError
If the preprocessor is not fitted.
"""
check_is_fitted(self, ['w', 'b'])
def _check_is_fitted(self):
"""Check if the classifier is trained (fitted).
Raises
------
NotFittedError
If the classifier is not fitted.
"""
# Do not check `b` as some classifiers do not set it
check_is_fitted(self, 'w')
super(CClassifierLinear, self)._check_is_fitted()
def _check_is_fitted(self):
"""Check if the classifier is trained (fitted).
Raises
------
NotFittedError
If the classifier is not fitted.
"""
if self._kernel is not None:
check_is_fitted(self, '_tr')
super(CClassifierRidge, self)._check_is_fitted()
def _check_is_fitted(self):
"""Check if the classifier is trained (fitted).
Raises
------
NotFittedError
If the classifier is not fitted.
"""
if not self.is_kernel_linear():
check_is_fitted(self, '_tr')
super(CClassifierSGD, self)._check_is_fitted()
def _check_is_fitted(self):
"""Check if the classifier is trained (fitted).
Raises
------
NotFittedError
If the classifier is not fitted.
"""
if not self.is_kernel_linear() or self.store_dual_vars is True:
check_is_fitted(self, 'sv') # Checking the SVs is enough
# SVM is a special case, is not set '_w' if kernel is not linear
# so we cannot call the superclass `_check_is_fitted`
if self.is_kernel_linear():
check_is_fitted(self, 'w')
# Then check the attributes of CClassifier
check_is_fitted(self, ['classes', 'n_features'])