patternnet algorithm for classification

Question

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

Kshitij Singh · Accepted Answer

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.

patternnet algorithm for classification

Get a Free Consultation or a Sample Assignment Review!