Skip to content
Open access

Multi-Algorithm Optimization of Explanation Stability in Ensemble Learning for Human Capital Attrition Risk Prediction

Sep 2026 · Mathematics · Vol 14, pp. 3283 · 3 citations · 41 references

TL;DR

Hyperparameter optimization for an ensemble of gradient-boosted and bagged learners becomes a bi-objective program maximizing the cost-sensitive area under the precision–recall curve jointly with the expected agreement of SHapley Additive exPlanations (SHAP) global rankings across bootstrap refits.

Abstract

Attrition models are tuned for discrimination, but the decision they support is a ranking of retention drivers, and nothing requires that ranking to survive retraining. We treat its reproducibility as an objective: hyperparameter optimization for an ensemble of gradient-boosted and bagged learners becomes a bi-objective program maximizing the cost-sensitive area under the precision–recall curve jointly with the expected agreement of SHapley Additive exPlanations (SHAP) global rankings across bootstrap refits. Four results follow, three specializing standard theory. Hoeffding’s U-statistic theory makes the estimator unbiased with variance 4ζ1/R+O(R−2), turning “how many refits” into a variance calculation; an elementary spacing argument makes a ranking’s survival a signal-to-noise rather than a noise-magnitude quantity; SHAP linearity makes the retraining variance in the ensemble attributions a quadratic form, so the mixing weights solve a minimum-variance portfolio. The fourth delimits the third: a ranking is invariant to attribution scale, and this variance is not, so attribution variance is no proxy for explanation stability—minimizing it selects the least stable learner on every benchmark—and stability is optimized directly. The cost of the scheme is given in closed form. Our central empirical finding is a caution. Across five public benchmarks spanning 1.4–50.6% positives, no test separates the multi-objective solvers from the single-objective ones at a practitioner’s budget, and the objective must be estimated to be optimized: at the refit budget such a search would choose, the estimator’s standard deviation exceeds the spread of true stability across the configurations it ranks, so maximizing the estimate overfits it. Re-estimating each search’s own selection with draws that took no part in selecting it leaves a resolvable stability gain on two of the five benchmarks. Giving the search five and 12.5 times the budget surfaces gains that survive where the practitioner’s budget surfaces none, so part of what a cheap search misses is real; but the gap between what a search advertises and what survives does not close as the budget grows, and on one benchmark widens. Declaring the objective is worth doing; what it pays cannot be read off the search that declared it. The caution applies to any model selection that resamples to estimate its criterion and then maximizes it.

Read PDF

Similar papers

Open access Nov 2026

MAC-Stack: Calibrated Multi-Model Ensemble Learning with Exact SHAP Attribution for Explainable Financial Distress Prediction

Multi-model ensembles of gradient-boosted trees are among the most accurate tools for financial distress prediction, but they are usually deployed as three loosely coupled steps: ad hoc fusion weights, post-hoc probability calibration and SHAP explanations of individual members. This paper proposes MAC-Stack (margin-ad...

Mira Lam, Caroline Sullivan · 0 citations
Book Open access Nov 2026

Diversification Across Predictors: Heterogeneous Ensemble Learning for Cross-Sectional Return Prediction

Cross-sectional stock return prediction remains challenging due to extremely low signal-to-noise ratios, high-dimensional factor spaces, and non-stationary predictor effectiveness. Individual model classes exhibit complementary failure modes in this regime: neural networks overfit noise due to high variance, while line...

Dan-Ning Li · 0 citations
Open access Sep 2026

Enhancing Transparency in Income Prediction: A SHAP-based Explainable AI Approach with Random Forest

Explainable Artificial Intelligence (XAI) has gained increasing importance as modern machine learning models operate as opaque black-box systems, limiting transparency and trust in automated decision-making. While most XAI frameworks merely focus on static post-hoc visualizations, this study goes beyond passive diagnos...

Marsani Asfi, B. Warsito, Adi Wibowo · 0 citations
Open access Nov 2026

From Prediction to Decision: Leakage-Controlled Machine Learning with Uncertainty-Calibrated Black-Litterman Allocation

A leakage-controlled decision-support pipeline combining weekly Elastic Net and Random Forest excess-return forecasts with uncertainty-calibrated Black–Litterman allocation across five ETFs is evaluated, finding that forecast ranking and portfolio ranking can diverge, but the available holdout does not establish statis...

Nicola Zanfardino, Marco Giovannini, L. Ricciardi Celsi · 0 citations

A Hierarchy of Entropy-Shapley Games for Multivariate Predictive Uncertainty

A hierarchy of three entropy-based Shapley games makes this output-side choice explicit for any ordered multivariate outcome, ranging from per-component marginal entropy to fully joint entropy, and establishes a chain-rule decomposition of the joint attribution and characterize the cross-component term through conditio...

Niklas Koenen, Claudia Battistin, J. Van den Abeele et al. · 0 citations
Open access Oct 2026

Tail-risk-constrained portfolio optimization with cluster pre-selection and machine learning: risk-adjusted evidence from the S&P 500

Standard Gaussian risk frameworks underestimate extreme joint losses in large equity universes, exposing investors to unbudgeted drawdowns. This article develops and tests a portfolio-construction pipeline that disciplines tail risk while exploiting forward-looking return signals. The four-stage pipeline int...

David Toscano, Juan C. Roca · 0 citations

We use cookies to run the site and, with your consent, for analytics and to show ads. See our Cookie Policy.