Benchmarking Symbolic Regression Methods

This project focuses on benchmarking modern symbolic regression methods in comparison to other common machine learning methods. This benchmark consists of more than 100 datasets from PMLB.

v1.0 was reported in our GECCO 2018 paper:

Orzechowski, P., La Cava, W., & Moore, J. H. (2018). Where are we now? A large benchmark study of recent symbolic regression methods. GECCO 2018. DOI, Preprint

How to run

Batch jobs are controlled via submit_jobs.py. Run python submit_jobs.py -h to see options.

Results of single methods on datasets are generated using analyze.py. Run python analyze.py -h to see options.

Contact

William La Cava (@lacava), lacava at upenn dot edu

Patryk Orzechowski (@athril), patryk dot orzechowski at gmail dot com