def AgentPredictorWorker.transform(self, data) [source]
Applies the champion model to generate P(class=1) scores.
Verifies that all required features are present in the
input to prevent silent failures in production.
Execute production inference: ML-ready data | → | probabilities. |
Parameters
data : pDF
ML-ready data (WOE-encoded for GLM, raw for GBM).
Must contain all columns in
Must contain all columns in
self.cols.selected.Returns
AgentPredictorResult
Predictions + full feature lineage metadata.
Raises
ValueError: If any required feature is missing from the input.