SPIDER The Spider Objects

Kernel Partial Least Squares by Rosipal et al.

   A=kpls(C) returns a kernel partial least squares  object
    object initialized with hyperparameters H.  
    Hyperparameters, and their defaults
     nroflatent=5;	 maximum number of latent variables. depends also on
     epsilon          stopping criteria - determines number of latent
     epsilon2         stopping criteria for found directions
     a.dtraining=[]; data used for training 
     a.Kc=[];        centered kernel matrix
     a.U=[];         found output directions
     a.T=[];         found input directions
     child 		 -- kernel 
     original;     -- 
     train, test

Reference : Kernel Partial Least Squares Regression in Reproducing Kernel Hilbert Space
Author : Roman Rosipal , Leonard J. Trejo
Link : http://www.kernel-machines.org/jmlr/