[FEATURE REQUEST] Support for regression based problems

AI4LIFE-GROUP / OpenXAI

OpenXAI : Towards a Transparent Evaluation of Model Explanations

https://open-xai.github.io/

MIT License

227 stars 37 forks source link

[FEATURE REQUEST] Support for regression based problems #13

Closed tabularaza27 closed 10 months ago

tabularaza27 commented 1 year ago

Are you planning to add support for regression tasks ? If so, I'd suggest replacing the exact equality condition of the original and perturbed input predictions in the stability metrics (ŷ_x == ŷ_x') to something like |ŷ_x- ŷ_x'|< eps, where eps is a user-defined range.

ehsankiakojouri commented 1 year ago

For Ground_truth evaluations, one could use benchmark functions used in optimization like Ackley (or any other known mathematical functions) and fit several linear regression models for several particular ranges of inputs. Weights of linear regression models could be used as ground truth for faithfulness metrics.