def test_select_univariate(self):
"""
Suppose the data follows a bimodal distribution. The model selector should be able to
figure out that the GaussianKDE is best.
"""
model = select_univariate(self.bimodal_data, self.candidates)
assert isinstance(model, GaussianKDE)
def test_truncated(self):
"""
Suppose the data follows a truncated normal distribution. The KS statistic should be
larger for a Gaussian model than a TruncatedGaussian model (since the fit is worse).
"""
model = select_univariate(
self.truncated_data,
[GaussianUnivariate(), TruncatedGaussian()])
assert isinstance(model, TruncatedGaussian)
def test_binary(self):
"""
Suppose the data follows a Bernoulli distribution. The KS statistic should be larger
for a TruncatedGaussian model than a GaussianKDE model which can somewhat capture a
Bernoulli distribution as it resembles a bimodal distribution.
"""
model = select_univariate(
self.binary_data,
[GaussianKDE(), TruncatedGaussian()])
assert isinstance(model, GaussianKDE)
def fit(self, X):
"""Fit the model to a random variable.
Arguments:
X (numpy.ndarray):
Values of the random variable. It must have shape (n, 1).
"""
self._instance = select_univariate(X, self.candidates)
self._instance.fit(X)
self.fitted = True
def fit(self, X):
"""Fit the model to a random variable.
Arguments:
X (numpy.ndarray):
Values of the random variable. It must have shape (n, 1).
"""
if self.selection_sample_size and self.selection_sample_size < len(X):
selection_sample = np.random.choice(X, size=self.selection_sample_size)
else:
selection_sample = X
self._instance = select_univariate(selection_sample, self.candidates)
self._instance.fit(X)
self.fitted = True
def fit(self, X):
"""Fits the model.
Arguments:
X: `np.ndarray` of shape (n, 1).
Returns:
None
"""
self.constant_value = self._get_constant_value(X)
if self.constant_value is None:
self._instance = select_univariate(X, self.candidates)
self._instance.fit(X)
else:
self._replace_constant_methods()
self.fitted = True