Hello.
How should one handle feature selection (for predictors) when you have a feature, say revenue retention (%) that is made of other features including, upgrades(), downgrades (), cancellations($), etc. Should I use either the revenue retention and discard the rest (since they are all in the calculation of revenue retention). Does it make sense to include all 3 features (upgrades, downgrades, cancellations) AND also revenue retention in the model?
Thanks