This repository provides the supporting code for the Monte Carlo Bounds for Reasonable Predictions (MC-BRP) algorithm described in Why Does My Model Fail? Contrastive Local Explanations for Retail Forecasting in the proceedings of the ACM conference on Fairness, Accountabillity, and Transparency (FAT* 2020) in Barcelona.

Since the dataset used in the paper is private, we provide an analogous dataset and model about predicting the critical temperature of molecular compunds from the UCI machine learning repository: https://archive.ics.uci.edu/ml/datasets/Superconductivty+Data. MC-BRP is applicable to any regression task with numeric features. It can also be applied to classification tasks where all errors are defined as large errors.