Improving the Validity of Decision Trees as Explanations

TL;DR

This paper proposes a method to improve decision tree explanations by training shallow trees that minimize maximum leaf error, enhancing validity and fairness compared to traditional trees.

Contribution

Introduces a training approach for shallow decision trees that reduces unbalanced leaf accuracy, improving explanation validity and fairness.

Findings

Shallow trees with minimized maximum leaf error provide more valid explanations.

The proposed method achieves competitive accuracy with state-of-the-art models.

Extending leaves with additional models enhances overall performance.

Abstract

In classification and forecasting with tabular data, one often utilizes tree-based models. Those can be competitive with deep neural networks on tabular data and, under some conditions, explainable. The explainability depends on the depth of the tree and the accuracy in each leaf of the tree. We point out that decision trees containing leaves with unbalanced accuracy can provide misleading explanations. Low-accuracy leaves give less valid explanations, which could be interpreted as unfairness among subgroups utilizing these explanations. Here, we train a shallow tree with the objective of minimizing the maximum misclassification error across all leaf nodes. The shallow tree provides a global explanation, while the overall statistical performance of the shallow tree can become comparable to state-of-the-art methods (e.g., well-tuned XGBoost) by extending the leaves with further models.