def _check_params(self):
if self.sample_support is None:
self.sample_support = self.views[0].shape[0]
self.sample_weights = np.ones(self.views[0].shape[0])
self.sample_weights /= np.linalg.norm(self.sample_weights)
self.c = _process_parameter("c", self.c, 2, len(self.views))
self.positive = _process_parameter("positive", self.positive, False,
len(self.views))
def _check_params(self):
self.kernel = _process_parameter("kernel", self.kernel, "linear",
self.n_views)
self.gamma = _process_parameter("gamma", self.gamma, None,
self.n_views)
self.coef0 = _process_parameter("coef0", self.coef0, 1, self.n_views)
self.degree = _process_parameter("degree", self.degree, 1,
self.n_views)
self.c = _process_parameter("c", self.c, 0, self.n_views)
def _check_params(self):
if self.c is None:
warnings.warn(
"c parameter not set. Setting to c=1 i.e. maximum regularisation of l1 norm"
)
self.c = _process_parameter("c", self.c, 1, len(self.views))
if any(c < 0 or c > 1 for c in self.c):
raise ValueError(
"All regularisation parameters should be between 0 and 1 "
f"1. c=[{self.c}]")
shape_sqrts = [np.sqrt(view.shape[1]) for view in self.views]
self.t = [max(1, x * y) for x, y in zip(self.c, shape_sqrts)]
self.positive = _process_parameter("positive", self.positive, False,
len(self.views))
def _check_params(self):
"""check number of views=2"""
if len(self.views) != 2:
raise ValueError(f"SpanCCA requires only 2 views")
cov = self.views[0].T @ self.views[1] / self.n
# Perform SVD on im and obtain individual matrices
P, D, Q = np.linalg.svd(cov, full_matrices=True)
self.P = P[:, :self.rank]
self.D = D[:self.rank]
self.Q = Q[:self.rank, :].T
self.max_obj = 0
if self.regularisation == "l0":
self.update = _support_soft_thresh
self.c = _process_parameter("c", self.c, 0, len(self.views))
elif self.regularisation == "l1":
self.update = _delta_search
self.c = _process_parameter("c", self.c, 0, len(self.views))
self.positive = _process_parameter("positive", self.positive, False,
len(self.views))
def _check_params(self):
self.c = _process_parameter("c", self.c, 0, len(self.views))
self.lam = _process_parameter("lam", self.lam, 1, len(self.views))
if self.mu is None:
self.mu = [
lam / np.linalg.norm(view)**2
for lam, view in zip(self.lam, self.views)
]
else:
self.mu = _process_parameter("mu", self.mu, 0, len(self.views))
self.eta = _process_parameter("eta", self.eta, 0, len(self.views))
if any(mu <= 0 for mu in self.mu):
raise ValueError("At least one mu is less than zero.")
_check_Parikh2014(self.mu, self.lam, self.views)
self.eta = [
np.ones(view.shape[0]) * eta
for view, eta in zip(self.views, self.eta)
]
self.z = [np.zeros(view.shape[0]) for view in self.views]
self.l1_ratio = [1] * len(self.views)
def _check_params(self):
self.c = _process_parameter("c", self.c, 0, len(self.views))
self.l1_ratio = _process_parameter("l1_ratio", self.l1_ratio, 0,
len(self.views))
self.positive = _process_parameter("positive", self.positive, False,
len(self.views))
self.regressors = []
for alpha, l1_ratio, positive in zip(self.c, self.l1_ratio,
self.positive):
if self.stochastic:
self.regressors.append(
SGDRegressor(
penalty="elasticnet",
alpha=alpha / len(self.views),
l1_ratio=l1_ratio,
fit_intercept=False,
tol=self.tol,
warm_start=True,
random_state=self.random_state,
))
elif alpha == 0:
self.regressors.append(
Ridge(
alpha=self.tol,
fit_intercept=False,
positive=positive,
))
else:
self.regressors.append(
ElasticNet(
alpha=alpha / len(self.views),
l1_ratio=l1_ratio,
fit_intercept=False,
warm_start=True,
positive=positive,
random_state=self.random_state,
))
def _check_params(self):
self.c = _process_parameter("c", self.c, 0.0001, len(self.views))
if any(c <= 0 for c in self.c):
raise ("All regularisation parameters should be above 0. "
f"c=[{self.c}]")
def _check_params(self):
self.nearest_neighbors = _process_parameter(
"nearest_neighbors", self.nearest_neighbors, 1, self.n_views
)
self.gamma = _process_parameter("gamma", self.gamma, None, self.n_views)
self.kernel = _process_parameter("kernel", None, "rbf", self.n_views)
def _check_params(self):
self.c = _process_parameter("c", self.c, 0, self.n_views)
def _check_params(self):
self.c = _process_parameter("c", self.c, 0, self.n_views)
self.view_weights = _process_parameter(
"view_weights", self.view_weights, 1, self.n_views
)
