def normal(self, X, y):
"""Use normal equation regularized to compute the parameters."""
X = features_reshape(X)
X = np.concatenate((np.ones((X.shape[0], 1)), X), axis=1)
L = np.identity(X.shape[0])
L[0, 0] = 0
self.optimizer.parameters = (
np.linalg.pinv(X.T.dot(X) + self.regularization * L).dot(X.T).dot(y)
)
def predict_probability(self, X):
"""Predict the probability of a target given features."""
if self.parameters is None or self.labels is None:
raise NotFitted("the model must be fitted before usage")
X = features_reshape(X)
if self.standardize is not None:
X = self.standardize(X)
X = np.concatenate((np.ones((X.shape[0], 1)), X), axis=1)
return self.hypothesis(X, self.parameters)
def fit(self, X, y):
"""Fit the model."""
X = features_reshape(X)
if not check_dataset_consistency(X, y):
raise InvalidInput("the features set and target set must have as many rows")
if self.standardize is not None:
X = self.standardize(X)
X = np.concatenate((np.ones((X.shape[0], 1)), X), axis=1)
self.optimizer(X, y)
def score_samples(self, X):
"""Compute the log-likelihood of all samples."""
if self._U_reduced is None or self._mean is None:
raise NotFitted("the model must be fitted before usage")
X = features_reshape(X)
n_features = X.shape[1]
precision = np.linalg.inv(self._covariance_matrix)
residuals = X - self._mean
return -(1 / 2) * (-np.log(np.linalg.det(precision)) + np.sum(
(residuals * np.dot(residuals, precision)), axis=1) +
n_features * np.log(2 * np.pi))
def fit(self, X):
"""Train the model."""
X = features_reshape(X)
max_dimension = min(X.shape)
if self.n_components is None:
self.n_components = max_dimension
elif self.n_components > max_dimension:
raise InvalidInput(
f"n_components must be lesser than {max_dimension}")
self._covariance_matrix = np.cov(X.T)
self._mean = np.mean(X, axis=0)
U, _, _ = np.linalg.svd(self._covariance_matrix)
self._U_reduced = U[:, :self.n_components]
def fit(self, X, y):
"""Fit the model."""
X = features_reshape(X)
if not check_dataset_consistency(X, y):
raise InvalidInput(
"the features set and target set must have as many rows")
if self.standardize is not None:
X = self.standardize(X)
X = np.concatenate((np.ones((X.shape[0], 1)), X), axis=1)
self.labels = np.unique(y)
n_labels = np.size(self.labels)
if n_labels < 2:
raise InvalidInput(
"target must have at least two different classes")
elif n_labels == 2:
self.optimizer(X, y)
else:
self.optimizer(X, (y == self.labels).astype(int))
def normal(self, X, y):
"""Use normal equation to compute the parameters."""
X = features_reshape(X)
X = np.concatenate((np.ones((X.shape[0], 1)), X), axis=1)
self.optimizer.parameters = np.linalg.pinv(X.T.dot(X)).dot(X.T).dot(y)