grnntra

 GRNNTRA Generalized Regression Neural Network Train

   grnn = grnntra(y,x,C,L,centers,order)

 INPUTS
       y: output training points (Ry,N)
       x: input  training points (Rx,N)
       C: kernel parameters (kernel widths)
       L: Weight decay regularization parameter
 centers: Model centers
   order: Sugeno model order


 Trains the GRNN and produces the following structure

     grnn.m: RBF centers
     grnn.C: RBF kernel parameters
     grnn.W: RBF weights
 grnn.order: Sugeno order of the approximation

 Other call conventions:

   grnn = grnntra(y,x)             Automatically computes kernel parameters and weights
   grnn = grnntra(y,x,C)           Original Spetch's GRNN algorithm (no least squares weights) 
   grnn = grnntra(y,x,C,L)         Uses least squares interpolation with regularization (weight decay factor L). 
   grnn = grnntra(y,x,L,'e')       Same as above but computes kernel parameters automatically
   grnn = grnntra(y,x,C,L,centres) Uses 'centres' instead of 'x'

 See also GRNNSIM