SPIDER The Spider Objects

L1 Norm Minimization for kernel classifiers.

    Minimize ||\alpha||_norm describing kernel expansion, under 
    separability constraints. [currently only for pattern recognition and
    norms 0, 1 and 2]
   A=ANORM(H) returns an anorm object initialized with hyperparameters H. 
   Hyperparameters, and their defaults:
    norm=1               -- norm to use, i.e. minimize ||\alpha||_norm
    ridge=1e-12          -- a ridge on the kernel
    child=kernel         -- the kernel is stored as a member called "child"
    alpha                -- the weights
    b0                   -- the threshold
    Xsv                  -- the Support Vectors
    train, test
   Refernce  : Use of the Zero-Norm With Linear Models and Kernel Methods  
Author : Jason Weston, André Elisseeff, Bernd Schölkopf, Mike Tipping
Link : http://citeseer.ist.psu.edu/531078.html