Estimation of interventional effects of features on prediction

TL;DR

This paper introduces a framework linking causal structures of data generation and prediction mechanisms, enabling estimation of feature interventions to achieve desired predictions, with a focus on linear data.

Contribution

It proposes a novel framework that identifies key causal features influencing predictions and estimates interventions, bridging causal inference and predictive modeling.

Findings

Framework successfully identifies influential features in artificial data.

Framework estimates effective interventions for desired predictions.

Demonstrated applicability on real-world datasets.

Abstract

The interpretability of prediction mechanisms with respect to the underlying prediction problem is often unclear. While several studies have focused on developing prediction models with meaningful parameters, the causal relationships between the predictors and the actual prediction have not been considered. Here, we connect the underlying causal structure of a data generation process and the causal structure of a prediction mechanism. To achieve this, we propose a framework that identifies the feature with the greatest causal influence on the prediction and estimates the necessary causal intervention of a feature such that a desired prediction is obtained. The general concept of the framework has no restrictions regarding data linearity; however, we focus on an implementation for linear data here. The framework applicability is evaluated using artificial data and demonstrated using real-world data.