Closed ArturoAmorQ closed 1 year ago
@lfarhi noticed a degrade of the performance of the tree model at the end of wrap-up quiz M5. The reason is that Pandas>2.0 converts None into NaN, which broke the pipeline with ordinal encoding. This PR fixes the issue.
None
NaN
@lfarhi noticed a degrade of the performance of the tree model at the end of wrap-up quiz M5. The reason is that Pandas>2.0 converts
None
intoNaN
, which broke the pipeline with ordinal encoding. This PR fixes the issue.