patternnet algorithm for classification

Dear All;

I would like to know what is the algorithm of patternnet function or how it works to classification.

Answer 1

[Greg Heath]

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