3.8 Explanations

3.8.1 Row Explanations

3.8.2 Table Explanations

Explanations are generated by running ML_EXPLAIN_ROW or ML_EXPLAIN_TABLE on unlabeled data; that is, it must have the same feature columns as the data used to train the model but no target column.

Explanations help you understand which features have the most influence on a prediction. Feature importance is presented as a value ranging from -1 to 1. A positive value indicates that a feature contributed toward the prediction. A negative value indicates that the feature contributed toward a different prediction; for example, if a feature in a loan approval model with two possible predictions ('approve' and 'reject') has a negative value for an 'approve' prediction, that feature would have a positive value for a 'reject' prediction. A value of 0 or near 0 indicates that the feature value has no impact on the prediction to which it applies.

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.

After the ML_TRAIN routine, use the ML_EXPLAIN routine to train prediction explainers and model explainers for HeatWave AutoML. You must train prediction explainers in order to use ML_EXPLAIN_ROW and ML_EXPLAIN_TABLE. In earlier releases, the ML_TRAIN routine trains the default Permutation Importance model and prediction explainers. See Section 3.6, “Training Explainers”.