Skip to content

SKSurrogate is a suite of tools that implements surrogate optimization for expensive functions based on scikit-learn. The main purpose of SKSurrogate is to facilitate hyperparameter optimization for machine learning models and optimized pipeline design (AutoML).

License

Notifications You must be signed in to change notification settings

mghasemi/sksurrogate

Repository files navigation

Quality Travis Documentation Status Binder Licence

Logo

SKSurrogate

SKSurrogate is a suite of tools which implements surrogate optimization for expensive functions based on scikit-learn. The main purpose of SKSurrogate is to facilitate hyperparameter optimization for machine learning models and optimized pipeline design (AutoML).

The version of the surrogate optimization implemented here heavily relies on regressors. A custom regressor based on Hilbert Space techniques is implemented, but all scikit-learn regressors are accepted for optimization.

Finding an optimized pipeline -based on a given list of transformers and estimators is a time-consuming task. A version of evolutionary optimization has been implemented to reduce its time in lieu of global optimality.

Dependencies

SKSurrogate heavily depends on NumPy, scipy, and scikit-learn, for its main functionalities. Other dependencies include pandas, matplotlib, SALib, peewee, and tqdm.

Documentation

The documentation is produced by Sphinx and is intended to cover code usage as well as a bit of theory to explain each method briefly. For more details refer to the documentation at sksurrogate.rtfd.io.

License

This code is distributed under MIT license:

MIT License

Copyright (c) 2019 Mehdi Ghasemi

Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions:

The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software.

THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.

About

SKSurrogate is a suite of tools that implements surrogate optimization for expensive functions based on scikit-learn. The main purpose of SKSurrogate is to facilitate hyperparameter optimization for machine learning models and optimized pipeline design (AutoML).

Topics

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages