Learning Ensembles of Interpretable Simple Structure
This work addresses the need for transparent decision support in operations research and other domains where model interpretability is crucial, offering an incremental improvement over existing interpretable models by handling intricate feature interactions through data partitioning.
The paper tackles the problem of balancing predictive accuracy and interpretability in machine learning models for decision-making by proposing an algorithm that partitions data into subgroups with simplified feature interactions, enabling the use of simple interpretable models within each subgroup. The result is improved explainability and predictive accuracy, with decision boundaries that are more interpretable and aligned with domain intuition compared to global models.
Decision-making in complex systems often relies on machine learning models, yet highly accurate models such as XGBoost and neural networks can obscure the reasoning behind their predictions. In operations research applications, understanding how a decision is made is often as crucial as the decision itself. Traditional interpretable models, such as decision trees and logistic regression, provide transparency but may struggle with datasets containing intricate feature interactions. However, complexity in decision-making stem from interactions that are only relevant within certain subsets of data. Within these subsets, feature interactions may be simplified, forming simple structures where simple interpretable models can perform effectively. We propose a bottom-up simple structure-identifying algorithm that partitions data into interpretable subgroups known as simple structure, where feature interactions are minimized, allowing simple models to be trained within each subgroup. We demonstrate the robustness of the algorithm on synthetic data and show that the decision boundaries derived from simple structures are more interpretable and aligned with the intuition of the domain than those learned from a global model. By improving both explainability and predictive accuracy, our approach provides a principled framework for decision support in applications where model transparency is essential.