Generalized Optimal Classification Trees: A Mixed-Integer Programming Approach

Jiancheng Tu a {}^{\text{a}}, Wenqi Fan a, b {}^{\text{a, b}}, Zhibin Wu c {}^{\text{c}} · 2026

Read Generalized Optimal Classification Trees: A Mixed-Integer Programming Approach by Jiancheng Tu a {}^{\text{a}}, Wenqi Fan a, b {}^{\text{a, b}}, Zhibin Wu c {}^{\text{c}} (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2602-02173.txt.

Description

Jiancheng Tu and others propose a mixed-integer programming method for learning globally optimal classification trees under nonlinear metrics such as the F1 score. Optimal trees are prized for clarity. Global tuning has only recently become real on real data, and standard methods handle class imbalance poorly. The method pairs flexible mixed-integer models with a tailored branch-and-cut method, instance reduction, and warm-start plans. On fifty benchmark data, it optimizes nonlinear metrics efficiently with strong predictive results and shorter answer times than current methods.

Catalog record

CREATORJiancheng Tu a {}^{\text{a}}, Wenqi Fan a, b {}^{\text{a, b}}, Zhibin Wu c {}^{\text{c}}
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE80334
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2602.02173

Browse the full library · About the Tower of Babel