def transform(self, Kt, k=None):
"""Embed Kt into the `k` dimensional subspace.
:Parameters:
Kt : 1d or 2d array_like object
precomputed test kernel matrix. (if kernel=None);
test data in input space (if kernel is a Kernel object).
"""
if self._coeff is None:
raise ValueError("no KPCA computed")
if k == None:
k = self._coeff.shape[1]
if k < 1 or k > self._coeff.shape[1]:
raise ValueError("k must be in [1, %d] or None" % \
self._coeff.shape[1])
Ktarr = np.asarray(Kt, dtype=np.float)
if self._kernel is not None:
Ktarr = self._kernel.kernel(Ktarr, self._x)
Ktarr = kernel.kernel_center(Ktarr, self._K)
try:
return np.dot(Ktarr, self._coeff[:, :k])
except:
raise ValueError("Kt, coeff: shape mismatch")
def transform(self, Kt, k=None):
"""Embed Kt into the `k` dimensional subspace.
:Parameters:
Kt : 1d or 2d array_like object
precomputed test kernel matrix. (if kernel=None);
test data in input space (if kernel is a Kernel object).
"""
if self._coeff is None:
raise ValueError("no KPCA computed")
if k == None:
k = self._coeff.shape[1]
if k < 1 or k > self._coeff.shape[1]:
raise ValueError("k must be in [1, %d] or None" % \
self._coeff.shape[1])
Ktarr = np.asarray(Kt, dtype=np.float)
if self._kernel is not None:
Ktarr = self._kernel.kernel(Ktarr, self._x)
Ktarr = kernel.kernel_center(Ktarr, self._K)
try:
return np.dot(Ktarr, self._coeff[:, :k])
except:
raise ValueError("Kt, coeff: shape mismatch")
def learn(self, K):
"""Compute the kernel principal component coefficients.
:Parameters:
K: 2d array_like object
precomputed training kernel matrix (if kernel=None);
training data in input space (if kernel is a Kernel object)
"""
Karr = np.asarray(K, dtype=np.float)
if Karr.ndim != 2:
raise ValueError("K must be a 2d array_like object")
if self._kernel is None:
if Karr.shape[0] != Karr.shape[1]:
raise ValueError("K must be a square matrix")
else:
self._x = Karr.copy()
Karr = self._kernel.kernel(Karr, Karr)
self._K = Karr.copy()
Karr = kernel.kernel_center(Karr, Karr)
self._coeff, self._evals = kpca(Karr)
def learn(self, K):
"""Compute the kernel principal component coefficients.
:Parameters:
K: 2d array_like object
precomputed training kernel matrix (if kernel=None);
training data in input space (if kernel is a Kernel object)
"""
Karr = np.asarray(K, dtype=np.float)
if Karr.ndim != 2:
raise ValueError("K must be a 2d array_like object")
if self._kernel is None:
if Karr.shape[0] != Karr.shape[1]:
raise ValueError("K must be a square matrix")
else:
self._x = Karr.copy()
Karr = self._kernel.kernel(Karr, Karr)
self._K = Karr.copy()
Karr = kernel.kernel_center(Karr, Karr)
self._coeff, self._evals = kpca(Karr)