Sklearn PipelineInstall

Summary

Create a scikit-learn pipeline from configured transformers and estimators, or apply a saved pipeline to data to transform or predict.

Inputs

• pipelineData (required) — JSON string containing the trained pipeline
• predictFeatureColumns (required) — Comma-separated list of feature column names
• pythonPath — Path to Python executable with scikit-learn installed

Output shape

a JSON object containing the serialized pipeline, metadata (steps, has_estimator, feature_columns) and optionally a score

The score field is included only when the operation is used for training; otherwise the output includes only pipeline data and transformation results.

Examples

Example 1: Train a new pipeline

Set operation=train, provide featureColumns, targetColumn, preprocessors, estimator, and optionally pcaComponents or polyDegree; pipelineData is empty.

Example 2: Use a trained pipeline for prediction

Set operation=predict, provide pipelineData (base64 of trained pipeline) and predictFeatureColumns; optional pythonPath can be left at its default.