De-Normalizer

Node Development
1

Hello,

I wonder if there is no de-normalizer in KNIME. Since if I normalize data for a certain learner and then I want to predict some values, I need to scale it from normalized to non-normalized, i.e. original ranged values. Currently I do it manually by calculating min and max or mean and standard deviation. It would be
really helpful if a normalizer could result in a normalizing model which I can applied to automatic de-normalization.

Is it possible to implement this or do I mis-interpretate how to work with normalization?

One more question: Is there a module to calculate the Chi-Square error of two columns?

Thanks in advance,

Michael

2

This is a good idea. I am looking for a way to denormalize too...

Thank you...

3

I agree too, this would be very useful.

Simon.

4

There is an alternative solution in this post :

http://tech.knime.org/forum/knime-general/post-processing-mlp-predictions