def kernelize(self, x1, x2):
'''
:params: X: NxD
'''
if self.kernel == 'linear':
return Kernels.linear(x1, x2)
elif self.kernel == 'rbf':
return Kernels.rbf(x1, x2, gamma=self.gamma)
elif self.kernel == 'sigmoid':
return Kernels.sigmoid(x1, x2, gamma=self.gamma)
elif self.kernel == 'polynomial':
return Kernels.polynomial(x1, x2, d=self.d)
elif self.kernel == 'cosine':
return Kernels.cosine(x1, x2)
elif self.kernel == 'correlation':
return Kernels.correlation(x1, x2, gamma=self.gamma)
elif self.kernel == 'linrbf':
return Kernels.linrbf(x1, x2, gamma=self.gamma)
elif self.kernel == 'rbfpoly':
return Kernels.rbfpoly(x1, x2, d=self.d, gamma=self.gamma)
elif self.kernel == 'rbfcosine':
return Kernels.rbfpoly(x1, x2, d=self.d, gamma=self.gamma)
elif self.kernel == 'etakernel':
return Kernels.etakernel(x1, x2, d=self.d, gamma=self.gamma)
elif self.kernel == 'alignment':
return Kernels.alignment(x1, x2)
elif self.kernel == 'laplace':
return Kernels.laplacian(x1, x2, gamma=self.gamma)
elif self.kernel == 'locguass':
return Kernels.locguass(x1, x2, d=self.d, gamma=self.gamma)
elif self.kernel == 'chi':
return Kernels.chi(x1)
def kernelize(self, x1, x2, type = None):
'''
:params: x1: NxD
:params: x2: NxD
'''
if self.kernel == 'linear':
if not type:
return Kernels.linear_svdd(x1, x2)
else:
return Kernels.linear(x1, x2)
def kernelize(self, x1, x2):
'''
:params: X: NxD
'''
if self.kernel == 'linear':
return Kernels.linear(x1, x2)
elif self.kernel == 'rbf':
return Kernels.rbf(x1, x2)
elif self.kernel == 'sigmoid':
return Kernels.sigmoid(x1, x2)
elif self.kernel == 'polynomial':
return Kernels.polynomial(x1, x2)
elif self.kernel == 'cosine':
return Kernels.cosine(x1, x2)
elif self.kernel == 'correlation':
return Kernels.correlation(x1, x2)