Feature construction using explanations of individual predictions

TL;DR

This paper introduces a heuristic method called Explainable Feature Construction (EFC) that leverages explanations of individual predictions to efficiently generate meaningful features, improving model performance and interpretability.

Contribution

The paper presents a novel EFC approach that reduces feature search space by aggregating instance-based explanations, enabling faster and more effective feature construction.

Findings

EFC significantly reduces feature construction time across various operators.

EFC improves classification accuracy on multiple real-world datasets.

EFC produces interpretable features validated by domain experts.

Abstract

Feature construction can contribute to comprehensibility and performance of machine learning models. Unfortunately, it usually requires exhaustive search in the attribute space or time-consuming human involvement to generate meaningful features. We propose a novel heuristic approach for reducing the search space based on aggregation of instance-based explanations of predictive models. The proposed Explainable Feature Construction (EFC) methodology identifies groups of co-occurring attributes exposed by popular explanation methods, such as IME and SHAP. We empirically show that reducing the search to these groups significantly reduces the time of feature construction using logical, relational, Cartesian, numerical, and threshold num-of-N and X-of-N constructive operators. An analysis on 10 transparent synthetic datasets shows that EFC effectively identifies informative groups of attributes and constructs relevant features. Using 30 real-world classification datasets, we show significant improvements in classification accuracy for several classifiers and demonstrate the feasibility of the proposed feature construction even for large datasets. Finally, EFC generated interpretable features on a real-world problem from the financial industry, which were confirmed by a domain expert.