Sklearn PCAInstall

Summary

Transform principal component data back to original feature space.

Inputs

• pcaData (required) — JSON string containing the fitted PCA model.
• pcColumns (required) — Comma‑separated list of principal component column names to invert.
• pythonPath — Path to the Python executable that can run the scikit‑learn PCA script.

Output shape

list of records; each record retains original fields and adds reconstructed_ fields for each principal component.

Output length matches input length; reconstruction requires matching pcColumns with components used during fitting.

Examples

Example 1: Reconstruct original feature values from PCA components stored in a previous node.

Set PCA Model Data to the JSON output of a 'Fit Transform' step, PC Columns as 'PC1,PC2', and optionally specify Python Path if different from default.

Example 2: Apply inverse transform to a dataset with three feature columns.

pcaData = '{"components": [...], "explained_variance_ratio": [...]}'; pcColumns = 'PC1,PC2,PC3'; pythonPath = 'python3'.

Example 3: Inverse transform a single workflow item.

Input item contains feature columns 'f1,f2'; configure pcaData and pcColumns accordingly; the node outputs the same item with added fields 'reconstructed_f1' and 'reconstructed_f2'.