The Complete SPRLIB & ANNLIB

adapt_mlnet

- perform one learning cycle on a maximum likelihood network

SYNOPSIS

int adapt_mlnet (net, etap, alphap, etas, alphas, options, class)

ARGUMENTS

NET *net The NET to be trained.
double etap The learning rate for learning the positions of the kernels.
double alphap The momentum constant for learning the positions of the kernels.
double etas The learning rate for learning the variances of the kernels.
double alphas The momentum constant for learning the variances of the kernels.
long options Flags - see FUNCTION below.
int class The class number of the sample that was used to evaluate the network. This number indicates which part of the network is to be adapted.

RETURNS

TRUE if an error occured, FALSE otherwise.

FUNCTION

This function takes an evaluated net (evaluated with a sample of class class) and adapts the weights and biases of the units representing class class according the the maximum likelihood learning rule. The net->NetFlag is checked; it should be MLNET (see NET-flags. Then, all incoming weights of all output units are updated. Finally, all hidden units of class class are updated (note that the class is indicated by the (normally not used) InOutIndex field). The options flag is a combination of the HIST-flags and BP-flags. This function takes the same flags as bp_adapt_unit: BPACCUM (accumulate the delta's) and BPUPDATE (update the weights). BPBATCH is disabled, since this function is called once for each sample. The function also supposes that the network was initialized using the bp_init function.

NOTE

This routine is used by the learn_mlnet function.

SEE ALSO

learn_mlnet

This document was generated using api2html on Thu Mar 5 09:00:00 MET DST 1998