I would like to use MLP prediction.
The problem is that normalizing is mandatory... and after the MLP prediction, values are normalized too.
EDIT : I tried to add an image of my situation, but it doesn't work.
http://img3.imageshack.us/img3/4469/knime001.png
What can I do now to get back to the initial scale ?
How about joining the original data after the prediction?
Yeah, thank you !
You gave me some ideas : after joining the initial data, I can find the normalizing coefficients.
The Joiner node is in Data Manipulation > Column > Split & Combine > Joiner.
If you only use the min-max normalizing. You can denormalize also with the normalizer node.
Either you type your previous range into by hand.
Or you use the statistics node, transpose the table, filter the row (representing the column) being normalized, use the TableRow To Variable, and put the variables into the normalizer node(flowvariables).
This of course only works with one column to normalize. If you want to do this with multiple columns... I will attach an flow...
Hope that helps, Iris
OK, I'll try.
Regards.