PRTools contents

PRTools manual

userkernel

USERKERNEL

Construct user defined kernel mapping

   K = USERKERNEL(B,R,FUNC,P1,P2, ...)
   K = B*USERKERNEL([],R,FUNC,P1,P2, ...)
   W = USERKERNEL([],R,FUNC,P1,P2, ...)
   W = R*USERKERNEL([],[],FUNC,P1,P2, ...)
   K = B*W

Input
 R Dataset, representation set, default B
 B Dataset
 FUNC String with function name to compute kernels (proximities) by
 K = FEVAL(FUNC,B,R,P1,P2, ...)

Output
 W Trained kernel mapping
 K Kernel (proximity) matrix

Description

The kernel matrix K is computed according to the definition given by the  user supplied function FUNC. The size of K is [SIZE(B,1) SIZE(R,1)]

PRTools contents

PRTools manual