Altair RISE

A program to recognize and reward our most engaged community members

Nominate Yourself Now!

How to implement stacking and genetic algorithm

Hi i'm still new using rapid miner.
My situation is I don't know where to put genetic algorithm feature selection. I also want use k-cross to split the data. Here I attach what i have done so far. I'm not very sure is my flow is right or not.

Image: https://us.v-cdn.net/6030995/uploads/editor/ie/eanolxm0c2df.png

Image: https://us.v-cdn.net/6030995/uploads/editor/85/wyutxoqfu3o7.png

Image: https://us.v-cdn.net/6030995/uploads/editor/t2/tmxvhiqaz0tg.png

Image: https://us.v-cdn.net/6030995/uploads/editor/kn/jz36e92ac6it.png

Find more posts tagged with

AI Studio

Accepted answers

All comments

BalazsBaranyRM

Hi!

If you want to actually test the process of the feature selection, which in my experience the right way to do, then you would do the following:

Outer cross validation => feature selection => inner cross validation => learning model.

Of course this takes a lot of time but makes your modeling process completely validated and repeatable in the sense of "when you have new data in 4 months and a feature starts to become relevant, executing the process then will catch up with that and use the attribute".

If your goal is to analyze the data once and determine which features are relevant, then you can do it without the outer feature selection.

Regards,
Balázs