MATLAB: Problem with targets in nprtool

annDeep Learning Toolboxneural networknprtool

I'm trying to train a ANN using the nprtool but i can't select my targets. The targets option is empty while the inputs shows all my data. Why is it happening? Any idea?

Thanks in advance

Best Answer

If you have n classes ,then your target matrix for nprtool should be a n*m matrix where m is the number of training samples. Each column of the matrix should have a 1 in the row representing the corresponding class with the rest as 0.

Related Solutions

MATLAB: Patternnet algorithm for classification

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

MATLAB: ANN Tool in MATLAB

If there are c classes or categories and N examples of I-dimensional inputs.

 [ I N ] = size(input)
 [ c N  ] = size(target)

where the target matrix contains c-dimensional {0,1} unit vectors satisfying

 sum(target)           = ones(1,N)
 vec2ind(target)       = classindices
 ind2vec(classindices) = target

Look at the target vectors of the classification data examples in

 help nndatasets
 doc nndatasets