how "atestinput" could be recogrnised in matlb and how it is calculated in testing of image clasification؟

I am a little bit confused. "atestinput" is not defined its value like ineger or real. how it could be recogrnised in matlb. furthermore, can someone explain me how "atestinput" calculated in testing of image clasification.

    %Testingforward propagation 
    for i=1:numtestimages
        ytest=zeros(testoutputnodes,1);
        ytest(testlabels(i)+1,1)=1;
        for row=1:size(testimages,1)
            for columns=1:size(testimages,2)
                atestinput(columns+(row-1)*size(testimages,2),1)=testimages(row,columns,i);
            end
        end

[Andreas Goser]

In order to help you better: Can you share the code including example data so others can attempt to reproduce? And from your description, it is also not clear whether you are getting an error or if the results don't make sense.

Answer 1

[Image Analyst]

MATLAB will automatically create/define variables using the class of the first thing assigned to them.

By the way, I would not use ,1 with vectors. It's possible you might be confusing it to make it think you're using a 2-dimensional array.