Artifact for "GenTree: Using Decision Trees to Learn Interactions for Configurable Software"

Abstract

This document describes the artifact package accompanying the ICSE'21 paper "GenTree: Using Decision Trees to Learn Interactions for Configurable Software". The artifact includes GenTree source code, pre-built binaries, benchmark program specifications, and scripts to replicate the data presented in the paper. Furthermore, GenTree is applicable to new programs written in supported languages (C, C++, Python, Perl, Ocaml), or can be extended to support new languages easily. GenTree implementation is highly modular and optimized, hence, it can also be used as a framework for developing and testing new interaction inference algorithms. We hope the artifact will be useful for researchers who are interested in interaction learning, especially iterative and data-driven approaches.