How to identify if an ML model interpretation is affected by correlation and how to deal with it.
Thanks for the illuminating post, it helped materialize my scattered intutions :)
Sorry if the question is naive, but how do you "group" features for interpretation (e.g. by combining them in a linear combination like in a PCA ?) ?
Features that are in a group are shuffled, so that the values in a row always stay together.
See for example here: https://arxiv.org/abs/2104.11688
Thank you Christoph for this insightful post. Very useful..
very insightful, thank you
Thanks!
Thanks for the illuminating post, it helped materialize my scattered intutions :)
Sorry if the question is naive, but how do you "group" features for interpretation (e.g. by combining them in a linear combination like in a PCA ?) ?
Features that are in a group are shuffled, so that the values in a row always stay together.
See for example here: https://arxiv.org/abs/2104.11688
Thank you Christoph for this insightful post. Very useful..
very insightful, thank you
Thanks!