Internal function for normalizing the data. This function makes a linear transformation of the inputs such that the values lie between -1 and 1.
a vector or matrix that needs to be normalized.
If x is a vector, base is the minimum of x. If x is a matrix, base is a vector with the minimum for each of the columns of the matrix x.
if x is a vector, spread=base-max(x). If x is a matrix, spread is a vector calculated for each of the columns of x.
A vector or matrix with the resulting normalized values.
z=2*(x-base)/spread - 1