Our final layer will consist of a single neuron that, using the 32 inputs from the hidden layer, will predict a single output value for each observation.
Putting all the layers together, our MLP network structure will look like this:
for each observation.