search

microsoft / responsible-ai-toolbox-mitigations

Python library for implementing Responsible AI mitigations.
https://responsible-ai-toolbox-mitigations.readthedocs.io/en/latest/
MIT License
57 stars 6 forks source link

Rebalance class status progress message #42

Closed nushib closed 1 year ago

nushib commented 2 years ago

The Rebalance class status message contains a string related to imputation instead:

No columns specified for imputation. These columns have been automatically identified: [] Running oversampling...

mrfmendonca commented 1 year ago

This is actually an expected behavior. What happening here is that some classes have a default pre-processing pipeline. For the Rebalance class, for example, when using SMOTE (which doesn't work with categorical variables or missing values), the transform_pipe parameter must contain an encoder and an imputer. Otherwise, a default BasicImputer and EncoderOHE objects are created and inserted in the transform_pipe internal variable. When calling the rebalance method, the BasicImputer and EncoderOHE will run before running SMOTE. This message is printed by the BasicImputer created. You can suppress this message by setting verbose = False when instantiating the Rebalance class