[GSK-1078] Correlation detector#1178
Conversation
GSK-1078 Spurious correlation
Feature slices highly correlated with label
→ export independence test Find slices which are highly correlated to given labels (for classification models). |
|
|
||
| # Prepare dataset for slicing | ||
| df = dataset.df.copy() | ||
| df[dataset.target] = pd.Categorical(df[dataset.target]) |
There was a problem hiding this comment.
should we validate that a ds contains a target or it's done by the caller code already?
There was a problem hiding this comment.
Yes, currently the whole scan assumes that there is a target column, although I’m not sure there is a check. I will add that.
andreybavt
left a comment
There was a problem hiding this comment.
Generally looks good, if the target question isn't relevant feel free to merge
@andreybavt fixed a few things including the dataset target support in scan. |
|
Kudos, SonarCloud Quality Gate passed! |
No description provided.