outbrain / outrank

A Python library for efficient feature ranking and selection on sparse data sets.
https://dl.acm.org/doi/10.1145/3604915.3610636
BSD 3-Clause "New" or "Revised" License
19 stars 3 forks source link

Unified config interface for all algorithms #27

Open SkBlaz opened 1 year ago

SkBlaz commented 1 year ago

By just specifying the algorithm alongside its hyperparameters in a single struct, it should be simpler to explore different configurations (HPOs) for algorithms. Something along the lines of

--heuristic_config {"algorithm":` "MutualInformation", "executor": "numba", "hp_sparsity": 0.2, "hp_randomized": True}

Side effect - simplification of existing algo-decision process/parsing.

SkBlaz commented 10 months ago

@bmramor wdyt about this one, shall we get it operational?