The current feature_extractor primitives have a very limited functionality when it comes to detecting which features to process. This should be improved so that each one of them works specifically on their intended type of column.
Also add a boolean fixed keep argument to indicate whether the original columns should be kept or removed from X.
The current feature_extractor primitives have a very limited functionality when it comes to detecting which features to process. This should be improved so that each one of them works specifically on their intended type of column.
Also add a boolean fixed
keep
argument to indicate whether the original columns should be kept or removed fromX
.