Instance Set Specification
==========================

A fundamental part of machine learning from
a set of examples is the specification of 
which instances (examples) are to be used
to 'train' the learning procedure and which
instances will be used to 'test' how good 
the result is.

Update:
   This button updates to the current full set
   of instance names.

The frame consists of three subframes, the
total set of current instances, the instances
selected to be training instances and those 
selected to be test instances.

There are two ways in which the sets can be
determined.  

The first is through 'Percentage'.
A given percentage of the total set of (selected)
instances can be isolated to be the training set
and the rest will be part of the training set.
Which percentage should be used is determined by
the slider.

The second is through 'Selected' instances.  The
selected instances in the current set (the leftmost
frame) are transfered to the training or or test sets
as determined by the radio buttons, 'Train' or 'Test'.

