Impact Range Assessment (IRA): An Interpretable Sensitivity Measure for Regression Modelling

TL;DR

Impact Range Assessment (IRA) is a new interpretability method for regression models that quantifies the maximum potential influence of each predictor on the response variable, enhancing transparency and understanding.

Contribution

The paper introduces IRA, an interpretable sensitivity measure for regression models that effectively ranks predictor influence and is validated on synthetic and real datasets.

Findings

IRA distinguishes relevant from irrelevant predictors effectively.

IRA results are consistent across multiple evaluations.

Case study shows IRA improves model interpretability.

Abstract

While regression models capture the relationship between predictors and the response variable, they often lack intuitive accompanying methods to understand the influence of predictors on the outcome. To address this, we introduce an interpretability method called Impact Range Assessment (IRA), which quantifies the maximal influence of each predictor by measuring the total potential change in the response variable, across the predictor range. Validation using synthetic linear and nonlinear datasets demonstrates that relevant predictors produced higher IRA values than irrelevant ones. Moreover, repeated evaluations produced results closely aligned with those from the single-execution analysis, confirming the robustness of the method. A case study using a model that predicts pellet quality demonstrated that the IRA provides a simple and intuitive approach to interpret and rank predictor influence, thereby improving model transparency and reliability.