I’m training a neural network in matlab toolbox and I realize that perhaps normalizing the training data (input and output) could be affecting the results.
As an example, we have the input data x and output y of the neural network:

x=[611.6 1175 2231.48 3997.755  915.725]
y=[0.19 0.19 611.6 1786.45 2961.51 2909.11 611.6 0.19]

What would be the best way to normalize the data?
I’m using hyperbolic tangent sigmoid transfer function.