search

rodrigo-arenas / Sklearn-genetic-opt

ML hyperparameters tuning and features selection, using evolutionary algorithms.
https://sklearn-genetic-opt.readthedocs.io
MIT License
307 stars 77 forks source link

[FEATURE] Add more population-based nature-inspired algorithms #100

Closed firefly-cpp closed 2 years ago

firefly-cpp commented 2 years ago

Thanks for working on this exciting project! Is there a possibility to add more population-based nature-inspired algorithms to your framework?

Is there a possibility to call algorithm implementations from the NiaPy framework?

rodrigo-arenas commented 2 years ago

Hi @firefly-cpp, thanks for your message! At this moment, the package heavily relies on how DEAP structure the evolutionary algorithms, so at this point, it's possible to add more algorithms (there are already three variants), but only the ones that are compatible with DEAP

I've thought as well allowing other "backends" like NiaPy to use different algorithms from different frameworks, this would probably be version 1.0.0 of the package since it would require to re-write more than 70% of the library; at this point, I'm not really working on this 1.0.0 version, since it'd be worth it if the project gains more users/contributors to make it possible

I hope it makes sense, greetings

firefly-cpp commented 2 years ago

Thanks for your explanation.