DATASET_NORMALIZE_WORKERS¶
DATASET_NORMALIZE_FIT Worker¶
Overview¶
A new worker, DATASET_NORMALIZE_FIT, has been added to Workflows.
This worker standardizes numeric dataset columns using available (non-missing) values and generates the normalization parameters required for consistent data preprocessing. The resulting parameters can be used by downstream normalization and inverse-normalization workflows.
Key Features¶
- Added support for the
DATASET_NORMALIZE_FITworker in Workflows - Standardizes numeric dataset columns using non-missing values
- Computes and stores normalization parameters
- Supports consistent dataset preprocessing across workflows
- Integrates with normalization and inverse-normalization processes
DATASET_NORMALIZE_APPLY Worker¶
Overview¶
A new worker, DATASET_NORMALIZE_APPLY, is now available in Workflows.
This worker standardizes numeric dataset columns using normalization parameters generated by the DATASET_NORMALIZE_FIT worker. By applying previously calculated mean and standard deviation values, the worker ensures consistent normalization across datasets without recalculating the normalization parameters.
Key Features¶
- Added support for the
DATASET_NORMALIZE_APPLYworker in Workflows - Standardizes numeric dataset columns using precomputed normalization parameters
- Uses mean and standard deviation values generated by
DATASET_NORMALIZE_FIT - Ensures consistent normalization across multiple datasets
- Supports machine learning and analytics preprocessing workflows
DATASET_NORMALIZE_INVERSE Worker¶
Overview¶
A new worker, DATASET_NORMALIZE_INVERSE, is now available in Workflows.
This worker reverses a previously applied normalization transformation, restoring normalized numeric dataset columns back to their original scale using the normalization parameters generated by the DATASET_NORMALIZE_FIT worker.
The worker is particularly useful for converting model outputs, predictions, or transformed datasets back into their original units for interpretation, reporting, and downstream processing.
Key Features¶
- Added support for the
DATASET_NORMALIZE_INVERSEworker in Workflows - Restores normalized numeric columns to their original scale
- Uses normalization parameters generated by
DATASET_NORMALIZE_FIT - Supports inverse transformation without recalculating parameters
- Integrates seamlessly with normalization workflows