← Back to all news
arXiv Machine Learning By Koen M. F. Gorgels, Lasai Barre\~nada, Maarten van Smeden, Ben Van Calster, Ewout W. Steyerberg, Wouter A. C. van Amsterdam

Optimizing for the decision not the prediction: an exploration of Smooth Net Benefit as a training objective

Read the original on arXiv Machine Learning →

The study introduces Smooth Net Benefit (σNB), a differentiable approximation of Net Benefit, as a training objective aimed at aligning predictive models with threshold‑specific clinical decisions. Experiments on the Framingham cardiovascular risk dataset and 44 TabZilla datasets show that σNB training yields modest improvements for logistic regression but little to no benefit for more flexible models such as GAMs and XGBoost. The authors conclude that σNB is not a universal replacement for negative log‑likelihood training, though it may be worth exploring in contexts where model flexibility is limited.

Machine-generated by The Flow from the publisher's headline and feed description — not written or checked by a human. The full article lives at arXiv Machine Learning.

Related stories

arXiv Statistics ML
Oct 28

Cost-based variable selection for prediction models

The article introduces a decision‑analytic variable selection algorithm that optimizes clinical prediction models by balancing predictive performance against the costs and harms of measuring variables. It exhaustively compares net benefit across all predictor combinations, with a backward‑elimination extension for high‑dimensional data, and selects variables only when their added predictive value outweighs their measurement costs. The method, implemented in the R package NBvarsel, was demonstrated on simulated data and three real case studies, producing parsimonious models that maintained discrimination and calibration while reducing unnecessary clinical burden.

By Lasai Barre\~nada, Andrew Vickers, Laure Wynants, Bieke Tack, Dirk Timmerman, Ewout Steyerberg, Ben Van Calster
More like this →
arXiv Machine Learning
Oct 27

Tabular Foundation Models for Credit Risk Prediction: A Dedicated Review and Empirical Comparison

arXiv:2605.18147v3 Announce Type: replace Abstract: Predictive models are central to credit risk management, where the accuracy of default-probability and loss estimates affects the profitability of...

By Bart Baesens, Andreas Goethals, Stefan Lessmann, Simon De Vos, Cristi\'an Bravo, David Martens, Victor Medina-Olivares, Christophe Mues, Maria Oskarsd\'ottir, Seppe vanden Broucke, Tony Van Gestel, Tim Verdonck, Wouter Verbeke
llmsbenchmarks
More like this →
arXiv Machine Learning
5d ago

Leakage-Controlled Structure-Preserving Augmentation and Template-Aware Evaluation for Bangla Symptom-Based Disease Prediction

The study evaluates a leakage‑controlled, structure‑preserving augmentation method—train‑only asymmetric Bernoulli perturbation—on Bangla symptom‑based disease prediction. Using 725 unique profiles across 85 diseases and 172 binary symptoms, the authors compare this augmentation against duplication, oversampling, class weighting, and SMOTEN, finding that boosted‑tree models (XGBoost, LightGBM, CatBoost) gain substantial accuracy improvements while Bernoulli Naive Bayes shows limited benefit after statistical correction. A Hamming‑distance audit reveals that most held‑out profiles are very close to training data, and robustness tests confirm that XGBoost and LightGBM maintain gains even when one‑bit disease profiles are purged.

By Sabah Nushra, Shovan Samanta Turzo, Moinul Hossain, Md Kishor Morol
benchmarks
More like this →