Consider N I-dimensional input column vectors. Each is associated with 1 of c distinct classes and a corresponding c-dimensional {0,1} unit column vector obtained from a column of the unit matrix eye(c).
The corresponding dimensions of the input and target matrices are
[ I N ] = size(input)
[ c N ] = size(target)
The correspondence between the true class indices in {1:c} and c-dimensional unit target column vectors is via
target = full(ind2vec(trueclassindices))
trueclassindices = vec2ind(target)
For example
>> trueclassindices = [ 5 2 3 1 4 ]
>> target = full(ind2vec(trueclassindices))
yields
trueclassindices =
5 2 3 1 4
target =
0 0 0 1 0
0 1 0 0 0
0 0 1 0 0
0 0 0 0 1
1 0 0 0 0
PATTERNNET TRAINS the generic feedforward neural network FEEDFORWARDNET to map each input vector into it's corresponding target vector.
See the documentation
help patternnet
doc patternnet
and run the example.
Then search for some of my examples from BOTH the NEWSGROUP and ANSWERS
greg patternnet.
Hope this helps.
Thank you for formally accepting my answer
Greg
