Advice on regressor decomposition breakdown

[ggerogiokas]

Hi,

First of all thanks for the great work, on this library!

I am just wondering what the best way to decompose the contribution of each regressor to the overall regression component would be. I would also guess this also changes when you log the response vs not? What is the recommended way of decomposing this if we log the response and the parts are multiplicative in particular.

Many thanks, George