"linear regression"

ThomasPe86

Hello dear forum,

this is my first post :-)

i have a question about the linear regression. I use this tool with 25 parameters to get the target value. RapidMiner doesn't take all 25parameters, because not all are important to describe the target value.

Which methods does RapidMiner use to delete the non important parameters ? Correlation Coefficients ? average information content  (Entropie) ?

I'm sorry for my bad english.
Thanks in advance,

Thomas

haddock

Hi Thomas!

And welcome! As the help on this operator says, " It uses the Akaike criterion for model selection.", so I think the answer, at the bottom, is entropy. However... the source code says this is the work of one Ingo Merswa, who also stalks these corridors of lunacy, so I stand ready to be corrected!

Have fun!