Hi,
Can we get nice Normalization node that does a bunch of useful things.
Normalization for N inputs and equivalent outputs.
norm to total sum
1, 2, 3, 4, 5 => 0.2, 0.13 , 0.5, 0.26, 0.33 = 1
norm to current max value
1, 2, 3, 4, 5 => 0.2, 0.4 , 0.6, 0.8, 1.0 => 0.06, 0.13, 0.2, 0.26, 0.33 = 1.
or ie. norm by the amount of inputs :
1, 3, 4 , 5 => 0.25, 0.75, 1, 1.25 => 0.076, 0.23, 0.30, 0.38 = 1.
norm by length
...
Would help in neural networks and RBF solving, etc where the sum of weights are not equal to 1. would help to avoid cross fading artifacts etc.
also option to rescale the values on output to other value then 1.
support negative normalization.
thank you
Tristan
