def test_requires_probabilistic_classifier(self):
"""
Assert requires probabilistic classifier
"""
message = "requires a probabilistic binary classifier"
assert is_classifier(RadiusNeighborsClassifier)
assert not is_probabilistic(RadiusNeighborsClassifier)
with pytest.raises(yb.exceptions.YellowbrickError, match=message):
DiscriminationThreshold(RadiusNeighborsClassifier())
def test_requires_probabilistic_classifier(self):
"""
Assert requires probabilistic classifier
"""
message = "requires a probabilistic binary classifier"
assert is_classifier(RadiusNeighborsClassifier)
assert not is_probabilistic(RadiusNeighborsClassifier)
with pytest.raises(yb.exceptions.YellowbrickError, match=message):
DiscriminationThreshold(RadiusNeighborsClassifier())
def test_accepts_decision_function(self):
"""
Will accept classifiers with decision function
"""
model = LinearSVC
assert is_classifier(model)
assert is_probabilistic(model)
assert hasattr(model, "decision_function")
assert not hasattr(model, "predict_proba")
try:
DiscriminationThreshold(model())
except YellowbrickTypeError:
pytest.fail("did not accept decision function model")
def test_accepts_decision_function(self):
"""
Will accept classifiers with decision function
"""
model = LinearSVC
assert is_classifier(model)
assert is_probabilistic(model)
assert hasattr(model, "decision_function")
assert not hasattr(model, "predict_proba")
try:
DiscriminationThreshold(model())
except YellowbrickTypeError:
pytest.fail("did not accept decision function model")
def __init__(
self,
estimator,
ax=None,
n_trials=50,
cv=0.1,
fbeta=1.0,
argmax="fscore",
exclude=None,
quantiles=QUANTILES_MEDIAN_80,
random_state=None,
is_fitted="auto",
force_model=False,
**kwargs
):
# Perform some quick type checking to help users avoid error.
if not force_model and (
not is_classifier(estimator) or not is_probabilistic(estimator)
):
raise YellowbrickTypeError(
"{} requires a probabilistic binary classifier".format(
self.__class__.__name__
)
)
# Check the various inputs
self._check_quantiles(quantiles)
self._check_cv(cv)
self._check_exclude(exclude)
# Initialize the ModelVisualizer
super(DiscriminationThreshold, self).__init__(
estimator, ax=ax, is_fitted=is_fitted, **kwargs
)
# Set params
self.n_trials = n_trials
self.cv = cv
self.fbeta = fbeta
self.argmax = argmax
self.exclude = exclude
self.quantiles = quantiles
self.random_state = random_state
def test_requires_probabilistic_classifier(self):
"""
Assert requires probabilistic classifier
"""
message = "requires a probabilistic binary classifier"
class RoboClassifier(ClassifierMixin):
"""
Dummy Non-Probabilistic Classifier
"""
def fit(self, X, y):
self.classes_ = [0, 1]
return self
assert is_classifier(RoboClassifier)
assert not is_probabilistic(RoboClassifier)
with pytest.raises(yb.exceptions.YellowbrickError, match=message):
DiscriminationThreshold(RoboClassifier())
def __init__(self,
model,
ax=None,
n_trials=50,
cv=0.1,
fbeta=1.0,
argmax='fscore',
exclude=None,
quantiles=QUANTILES_MEDIAN_80,
random_state=None,
**kwargs):
# Perform some quick type checking to help users avoid error.
if not is_classifier(model) or not is_probabilistic(model):
raise YellowbrickTypeError(
"{} requires a probabilistic binary classifier".format(
self.__class__.__name__))
# Check the various inputs
self._check_quantiles(quantiles)
self._check_cv(cv)
self._check_exclude(exclude)
# Initialize the ModelVisualizer
super(DiscriminationThreshold, self).__init__(model, ax=ax, **kwargs)
# Set params
self.set_params(
n_trials=n_trials,
cv=cv,
fbeta=fbeta,
argmax=argmax,
exclude=exclude,
quantiles=quantiles,
random_state=random_state,
)
def __init__(self, model, ax=None, n_trials=50, cv=0.1, fbeta=1.0,
argmax='fscore', exclude=None, quantiles=QUANTILES_MEDIAN_80,
random_state=None, **kwargs):
# Perform some quick type checking to help users avoid error.
if not is_classifier(model) or not is_probabilistic(model):
raise YellowbrickTypeError(
"{} requires a probabilistic binary classifier".format(
self.__class__.__name__
))
# Check the various inputs
self._check_quantiles(quantiles)
self._check_cv(cv)
self._check_exclude(exclude)
# Initialize the ModelVisualizer
super(DiscriminationThreshold, self).__init__(model, ax=ax, **kwargs)
# Set params
self.set_params(
n_trials=n_trials, cv=cv, fbeta=fbeta, argmax=argmax,
exclude=exclude, quantiles=quantiles, random_state=random_state,
)
