Is it possible to create a CNN with real output?

The output type of the trainNetwork() must be categorical(). How can I create a CNN with float/real output(s)?

I mean the following command gives the following error:

>> convnet = trainNetwork(input_datas, [0.0, 0.1, 0.2, 0.3], networkLayers, opts);
Error using trainNetwork>iAssertCategoricalResponseVector (line 269)
Y must be a vector of categorical responses.

(The error message corresponds the [0.0, 0.1, 0.2, 0.3] vector), But I need real outputs, not categories.

The networkLayers is the following:

>> networkLayers= 

5x1 Layer array with layers:

     1   ''   Image Input       1x6000x1 images with 'zerocenter' normalization
     2   ''   Convolution       10 1x100 convolutions with stride [1  1] and padding [0  0]
     3   ''   Max Pooling       1x20 max pooling with stride [10  10] and padding [0  0]
     4   ''   Fully Connected   200 fully connected layer
     5   ''   Fully Connected   1 fully connected layer

Answer 1

[Sean de Wolski]

Use a regressionLayer at the end:

https://www.mathworks.com/help/releases/R2017a/nnet/ref/regressionlayer.html

Note this requires 17a, regression was added in that release.

Answer 2

[Mark Fajet]

Yes it is possible, however the documentation for TrainNetwork specifies that, for classification problems, that second parameter,"Y", must be "a categorical vector containing the image labels." A simple solution would be to write:

convnet = trainNetwork(input_datas, categorical([0.0, 0.1, 0.2, 0.3]), networkLayers, opts);

Later on, when using your neural network, if you really need the responses as an array of floats instead of a categorical array, you can convert a categorical array to an array of floats like this:

d = str2double(cellstr(c)); % Where c is a categorical array

Answer 3

[Greg Heath]

The simplest way is to represent the categories with integers

>> categories = [ 1 3 5 4 2]

>> target = full(ind2vec(categories))

   target     =   1     0     0     0     0
                  0     0     0     0     1
                  0     1     0     0     0
                  0     0     0     1     0
                  0     0     1     0     0

>> output = target + 0.1*randn(5,5)

   output =   0.8902   -0.1361   -0.0874    0.0327   -0.0846
             -0.1416    0.0780    0.0415   -0.0515    0.9827
              0.0060    1.0439    0.0348   -0.0896   -0.1209
             -0.0411   -0.0090    0.0349    0.8797   -0.0297
             -0.0368    0.1021    0.9271    0.1038   -0.3232

>> answer = vec2ind(output)

   answer = 1     3     5     4     2

Hope this helps.

 * Thank you for formally accepting my answer*

Greg