This is the module where the network learns the data patterns in the training set. Training continues until the network reaches the conditions set in the Training and Stop Training Criteria module. This module calls different learning subprograms depending upon the paradigm and architecture you select.
You have the option of viewing graphics and network training statistics as training progresses. (The selection varies with the type of network being used.) Training slows as more graphs or statistics are displayed.
Networks are sometimes sensitive to initial weight settings. If you want to try a set of random weights other than the default set when working with Backpropagation and Kohonen networks, you can choose a new random number seed before you start training by selecting the Run menu's Set Random Number Seed option.
File Note: This module defaults to training on a .TRN file, if it exists, or the .PAT file if there is no .TRN file. This module uses the .FIG file created in the Design module and the .MMX file created in the Define Inputs and Outputs module.
The Learning module can be minimized and run independently as training proceeds.