Final answer:
The mapping for a third-party predictive model's input fields is performed within the Predictive Model rule, which defines the correspondence between application data and model inputs.
Step-by-step explanation:
The mapping of the input fields of a third-party predictive model is done in the Predictive Model rule. When working with third-party predictive models in Pega systems, for instance, you define how the input variables from the Pega application should match with the expected inputs of the predictive model within the Predictive Model rule. This ensures that the data is correctly fed into the model for accurate predictions.