Effect sizes

[asheetal]

I want to model the effect sizes of predictors. My goal is that along with unexplained effects the sum total must be 100%. And then represent them (grouped) in a sort of pie chart.

Any guidance how should I approach this problem?

[hbaniecki]

Hi @asheetal, am I assuming correctly that you work primarily with statistical models? Reading through your problem, you might want to check out the effectsize package.

[asheetal]

Thanks @hbaniecki for your reply. I saw this package works with bayesian models as well. Can This work with xgboost models also?

[hbaniecki]

I have no clue, you might want to ask in the project's issues.