ROOT

Authors Yiren Hou, Peter Liu, Sean McGrath, Harsh Parikh

Last update 31 July 2026

Copyright license MIT license, File License

Software version identifier PACKAGES.rds, 0.1.1, 0.2.0

Description

Implements a general framework for globally optimizing user-specified objective functionals over interpretable binary weight functions represented as sparse decision trees, called ROOT (Rashomon Set of Optimal Trees). It searches over candidate trees to construct a Rashomon set of near-optimal solutions and derives a summary tree highlighting stable patterns in the optimized weights. ROOT includes a built-in generalizability mode for identifying subgroups in trial settings for transportability analyses (Parikh et al. (2025) <doi:10.1080/01621459.2025.2495319>).

Cites work

Who Are We Missing? : A Principled Approach to Characterizing the Underrepresented Population

This page was built for software: ROOT