Sklearn Robust Scaler icon

Sklearn Robust Scaler

Scale features using statistics robust to outliers (median and IQR)

Actions3

Transform

AI-generated

Summary

Transform data using a fitted robust scaler

Inputs

  • scalerData (required) — JSON string containing the fitted scaler model
  • transformColumns (required) — Comma-separated list of column names to transform
  • pythonPath — Path to the Python executable with scikit-learn installed (defaults to python3)

Output shape

Array of records, each containing original fields plus a new field per transformed column with the scaled value (named <column>_scaled)

Errors in column presence abort the workflow unless continueOnFail is enabled; the Python script must be executable and return JSON with transformed values

Examples

Example 1: Scale numeric columns col1 and col2 using a previously fitted scaler stored in scaler_json

"scalerData": "{"center_":[0.5],"scale_":[1.0],"with_centering":true,"with_scaling":true,"quantile_range":[25,75]},"transformColumns": "col1,col2","pythonPath": "python3"

Example 2: Inverse‑transform the scaled columns col1_scaled back to original space

"scalerData": "{"center_":[0.5],"scale_":[1.0]},"transformColumns": "col1_scaled","operation": "inverseTransform"

Discussion