The Conditioning Bias in Binary Decision Trees and Random Forests and Its Elimination
This addresses a subtle but impactful bias in widely used ML models, offering incremental improvements for practitioners relying on decision trees and random forests.
The paper tackles bias in binary decision trees and random forests due to conditioning operators, proposing techniques to eliminate it, resulting in statistically significant improvements of up to 0.1-0.2 percentage points in AUC and r² scores, with a 1.5 percentage point gain in the most sensitive case.
Decision tree and random forest classification and regression are some of the most widely used in machine learning approaches. Binary decision tree implementations commonly use conditioning in the form 'feature $\leq$ (or $<$) threshold', with the threshold being the midpoint between two observed feature values. In this paper, we investigate the bias introduced by the choice of conditioning operator (an intrinsic property of implementations) in the presence of features with lattice characteristics. We propose techniques to eliminate this bias, requiring an additional prediction with decision trees and incurring no cost for random forests. Using 20 classification and 20 regression datasets, we demonstrate that the bias can lead to statistically significant differences in terms of AUC and $r^2$ scores. The proposed techniques successfully mitigate the bias, compared to the worst-case scenario, statistically significant improvements of up to 0.1-0.2 percentage points of AUC and $r^2$ scores were achieved and the improvement of 1.5 percentage points of $r^2$ score was measured in the most sensitive case of random forest regression. The implementation of the study is available on GitHub at the following repository: \url{https://github.com/gykovacs/conditioning_bias}.