[Master Index] [Index for export/lyngby]



Function Synopsis

[E, Y, U] = lyngby_nnr_main(x, T, arg1, arg2, arg3, arg4, ...

Help text

 lyngby_nnr_main      - Main function for Neural network Regression

	function [E, Y, U] = lyngby_nnr_main(x, T, PropertyName, ...

       Input:  x   Neural network input (the paradigm)
               T   Target output (the datamatrix)
                  'GenOptim'      { {Free} | EarlyStop | 
                                  HiddenUnitsEarlyStop | Pruning |
                                  Pruning1DRegGridSearch |
                                  Pruning2DRegGridSearch ]
                                  Generalization optimization
                  'Validation'    [ {SingleBlocked}]
                                  How to compute the generalization
                  'Lag'           { 7 } Time lag.
                  'HiddenUnits'   { 3 } Number of hidden units, not
                                  counting the threshold unit
                  'Reg'           { 0.001 } Regularization parameter
                                  (weight decay)
                  'InputType'     [ {Direct} | SVDPreprocessed ] 
                  'SingularValues'   { 3 } Number of singular values
                                  maintained (in the case of SVD
                                  preprocessing), ie. the number of
                                  input neurons
                  'Info'          [ {0} | 1 ] Running information
                                  about the optimization process 

       Output: E   Error
               Y   Computed output
               U   Network weights

       Neural network regression is a generalization (with respect to
       non-linearity) compared to the FIR model.

       See also: lyngby_nn_qmain, lyngby_fir_main

Cross-Reference Information

This function calls
This function is called by

Produced by mat2html on Wed Jul 29 15:43:40 2009
Cross-Directory links are: OFF