SKSurrogate
sksurrogate/cite

Cite SKSurrogate

A repository link moves; a DOI does not. Cite the versioned software archive for the release you actually used, and add the dataset DOIs when your claim depends on the benchmark numbers.

MIT licence DOI 10.5281/zenodo.23631823 ORCID 0000-0002-9551-1780

The authoritative record is CITATION.cff at the tag you are citing — the entries below mirror it and are updated when a release is archived.

Software (BibLaTeX)

@software{ghasemi2026sksurrogate,
  author  = {Ghasemi, Mehdi},
  title   = {SKSurrogate: a Python toolkit for surrogate-based optimization
             of expensive functions},
  year    = {2026},
  version = {0.2.0},
  doi     = {10.5281/zenodo.23631823},
  url     = {https://github.com/mghasemi/sksurrogate},
  note    = {MIT licence}
}

Software (plain BibTeX)

Does your bibliography style not know @software? Use @misc — the fields are the same; only the entry type and the howpublished line differ.

@misc{ghasemi2026sksurrogate,
  author       = {Ghasemi, Mehdi},
  title        = {SKSurrogate: a Python toolkit for surrogate-based optimization
                  of expensive functions},
  year         = {2026},
  howpublished = {Zenodo software archive},
  version      = {0.2.0},
  doi          = {10.5281/zenodo.23631823},
  url          = {https://github.com/mghasemi/sksurrogate},
  note         = {MIT licence}
}

Datasets behind the published benchmarks

ArchiveContentsDOI
Equal-budget HPO suiteRaw B=30 / B=60 records, protocol DESIGN v1.910.5281/zenodo.23629024
D-track regime mapControlled synthetic study — raw data and report10.5281/zenodo.23652613
Software archiveSource release cited above10.5281/zenodo.23631823

In text

With the key above, a LaTeX citation is simply \cite{ghasemi2026sksurrogate}, or \parencite{ghasemi2026sksurrogate} under BibLaTeX.

If your work uses the benchmark protocol rather than the search layer, cite the dataset archive as well — the protocol contract it implements is benchmarks/DESIGN.md in the repository.

What to cite when the release has no DOI yet

Zenodo mints DOIs per uploaded release. If you are using a version that has not been archived yet, cite the most recent archived version in CITATION.cff and say so in your methods section; that keeps your citation resolvable and honest about which build you ran. The installed version is always available from the package itself: python -c "import sksurrogate; print(sksurrogate.__version__)".

© 2019–2026 Mehdi Ghasemi · MIT licence sksurrogate v0.3.0 · scikit-learn-native