When True, attach per-sample rule explanations to the response. Each rule is a small set of feature conditions that together characterize the model's prediction for that sample. When False, rules are skipped and the optional knobs below are inert.
max_rulesinteger nullable
Maximum number of rules returned per sample. Optional; the service default applies when unset.
max_depthinteger nullable
Maximum number of feature conditions allowed in a single rule. Optional; the service default applies when unset.
min_supportnumber nullable
Minimum fraction of the reference dataset a rule must match. Optional; the service default applies when unset.
max_features_consideredinteger nullable
Maximum number of input features available to the rule computation per sample. Optional; the service default applies when unset.