I’ve heard about a number of consulting companies, that decided to use simple linear model instead of a black box model with higher performance, because ,,client wants to understand factors that drive the prediction’’.
And usually the discussion goes as following: ,,We have tried LIME for our black-box model, it is great, but it is not working in our case’’, ,,Have you tried other explainers?’’, ,,What other explainers’’?
So here you have a map of different visual explanations for black-box models. Choose one in (on average) less than three simple steps.
These are available in the DALEX package. Feel free to propose other visual explainers that should be added to this map (and the package).
I wonder if anything different from outcome elasticities could be digested by business side, if regression models which treats variables in non-linead manner (like Generalized Additive Models) are unacceptable what could dispose clients to use coplots etc. Sometimes it seems to me that business side is really going to use model in way that isn’t explicitly stated – like reuseing estimated parameters is some processes in organization to optimize its performance (use sell prediction model for setting parameters of marketing campaign instead of optimizing campaign parameters directly) and we are not told about that because those are second order goals and the case of useing model is fuzzy for model owner at organisation itself.