A Granular Neural Network Simulator

Currently, a binary class problem must be formulated as a 1-of-2 code in
the GNN simulator. Otherwise, a confusion matrix cannot be generated.
However, 1-of-n codes are normally only used for three or more output
classes; a binary-class problem is normally formulated with a single output
variable with values of 0 or 1. The prupose of this RFE is to correct this
idiosyncracy in the GNN Simulator. The desired result is as follows: a
binary class problem can be formulated with a single output, and the
resulting confusion matrix is a proper 2x2 contingency table. Problems with
K classes must be formulated as a 1-of-K code, and the resulting confusion
matrix is a KxK contingency table. At this time, the size of the confusion
matrix is determined solely by the number of outputs.