Before you train a machine learning (ML) model, review the existing match models. If the existing match model configurations don't meet your business requirements, create a custom model.
The model that you identify might have pre-configured declarative match rules and candidate selection criteria. To create a hybrid match model that uses a combination of declarative match rules and an ML model, retain the declarative match rules. Edit the declarative match rules and the candidate selection criteria if required. To create a match model that uses an ML model only, delete any declarative match rules that the model contains.