SKSurrogate

SKSurrogate is a Python toolkit that simplifies hyperparameter tuning and pipeline optimization for machine learning models. It builds upon the popular scikit-learn library, making it familiar and easy to use.

Key Features:

• Effortless Hyperparameter Tuning: Efficiently find the best settings for your machine learning models.

• AutoML Optimization: Automate the process of designing optimal pipelines for your tasks (AutoML).

• Flexible Surrogate Optimization: Use various surrogate models, including a custom Hilbert Space regressor, or choose from scikit-learn's wide range of regressors for optimization.

• Faster Pipeline Design: Save time finding optimized pipelines through an implemented evolutionary optimization algorithm, even if it doesn't guarantee a globally optimal solution.

Documentation & Benchmarks:

• Documentation (install, API): sksurrogate.readthedocs.io · PyPI: pypi.org/project/sksurrogate

• An Honest Benchmark of My Own AutoML Package — equal-budget comparison vs. Optuna-TPE and random search; pre-registered protocol, 462 runs archived on Zenodo → link: https://sksurrogate.readthedocs.io/en/latest/blog-honest-benchmark.html

• Full benchmark tables + reproducibility provenance → link: https://sksurrogate.readthedocs.io/en/latest/benchmarks.html

• Software archive (v0.2.0): doi.org/10.5281/zenodo.23631823 · Raw data: doi.org/10.5281/zenodo.23629024

Overall, SKSurrogate empowers you to streamline the optimization process for your machine learning workflows.