[basis funcs] memory efficient posterior inference

This commit is contained in:
mzwiessele 2015-04-28 11:38:35 +02:00
parent be40318e0b
commit a24853da6b
2 changed files with 4 additions and 2 deletions

View file

@ -136,7 +136,7 @@ class DomainKernel(LinearSlopeBasisFuncKernel):
def _phi(self, X):
phi = np.where((X>self.start)*(X<self.stop), 1, 0)
return phi#((phi-self.start)/(self.stop-self.start))-.5
class LogisticBasisFuncKernel(BasisFuncKernel):
def __init__(self, input_dim, centers, variance=1., slope=1., active_dims=None, ARD=False, ARD_slope=True, name='logistic'):
self.centers = np.atleast_2d(centers)