Aliro icon indicating copy to clipboard operation
Aliro copied to clipboard

Add OneHotEncoder and OrdinalEncoder

Open jay-m-dev opened this issue 2 years ago • 0 comments

Add a OneHotEncoder and OrdinalEncoder. In a previous commit (# 9edbbce) these were removed. We should put back the Encoders in the validateDataset step to allow uploading datasets with ordinal and categorical columns.

This functionality was part of Aliro prior to the 9edbbce commit. But, experiments could not be run. It seems the Encoders were never implemented when running experiments. So, this issue will require 2 steps:

  1. Put back the encoders in the validateDataset step
  2. Perform encoding when a new experiment is launched.

Testing of this feature can be done with the datasets in data/datasets/test/integration/

Once this is implemented, we'll need to put back the skipped unittests in learn_tests.py and test_validateDataset.py

jay-m-dev avatar Jul 13 '23 20:07 jay-m-dev