Skip to content

Genetic algorithm for creating and optimizing a chain of sklearn models.

License

Notifications You must be signed in to change notification settings

skirklin/modopt

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

8 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

modopt

Toolkit for creating and optimizing an arbitrary ensemble of sklearn models.

sklearn only supports ensembles of a single classifier type, with a single set of parameters. I made this package to allow creating ensembles of multiple types of models, or the same model but with different model parameters. Since this represents a very large optimization space (model weights, model parameters), it is bundled with a simple genetic algorithm code to allow for global optimization over a very large parameter space.

Still under active development. Currently only an EnsembleRegressor is working, but an EnsembleClassifier and EnsembleCluster is coming. The genetic algorithm code is also working, but has not been integrated for actual model optimization yet. Work can also still be done to improve the performance of GA.

About

Genetic algorithm for creating and optimizing a chain of sklearn models.

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages