ClimbsRocks / auto_ml

[UNMAINTAINED] Automated machine learning for analytics & production
http://auto-ml.readthedocs.io
MIT License
1.64k stars 310 forks source link

let the user pass in their own model training code? #310

Open ClimbsRocks opened 7 years ago

ClimbsRocks commented 7 years ago

we expect it to take in a scipy sparse matrix, and return a trained model.

that trained model must then have a .predict (and ideally .predict_proba if a classifier) that takes in a scipy sparse matrix, and returns a list of predictions (one per prediction row).

this lets the user get crazy, and implement any algos they want.

combine this with .transform_only, and we're pretty covered.