Hi to all,
my problem is the following: I have 100 samples each one with dimension 3, and 100 corresponding outputs with dimension 1.
I would like to use the Deep Learning Toolbox, and in particular the trainNetwork function, for finding a function which given a certain input (of 3 elements) reproduces almost the correct corresponding single output (it should be a regression problem).
In the definition of the layers of the Neural Network, I am not sure if the input layer should be sequenceInputLayer(3) or featureInputLayer(3). Notice that my samples are not time depending: essentially, they are just 100 set of 3 random values taken between 0 and 1.
I think I need to use the featureInputLayer, but I am not sure why the sequenceInputLayer should not work.
Thanks in advanced for your help!