6.5.6 Generate Prediction Explanations

Prediction explanations are generated by running ML_EXPLAIN_ROW or ML_EXPLAIN_TABLE on unlabeled data. The data must have the same feature columns as the data used to train the model. The target column is not required.

Prediction explanations are similar to model explanations, but rather than explain the whole model, prediction explanations explain predictions for individual rows of data. See Explanations Overview to learn more.

You can train the following prediction explainers:

The Permutation Importance prediction explainer, specified as permutation_importance, is the default prediction explainer, which explains the prediction for a single row or table. Right after training and loading a model, you can run ML_EXPLAIN_ROW and ML_EXPLAIN_TABLE with this prediction explainer directly without having to run ML_EXPLAIN first.
The SHAP prediction explainer, specified as shap, uses feature importance values to explain the prediction for a single row or table. To run this prediction explainer with ML_EXPLAIN_ROW and ML_EXPLAIN_TABLE, you must run ML_EXPLAIN first.

ML_EXPLAIN_ROW generates explanations for one or more rows of data. ML_EXPLAIN_TABLE generates explanations on an entire table of data and saves the results to an output table. ML_EXPLAIN_* routines limit explanations to the 100 most relevant features.