Which task must be completed before EPM administrators import a Machine Learning model into Planning?

Before an EPM (Enterprise Performance Management) administrator can import a Machine Learning (ML) model into Oracle Planning, a prerequisite task must be completed by data scientists. According to Oracle’s "Bring Your Own ML" feature in the Planning application, the process begins with data scientists gathering historical data related to a business problem, training an ML algorithm, and generating a Predictive Model Markup Language (PMML) file using a third-party data science tool or Oracle Data Science Cloud. This PMML file represents a fully trained ML model that can then be imported into the Planning application by an EPM administrator.

Option A is correct because it aligns with this prerequisite step: the ML model must be pretrained and saved as a PMML file before the import process can begin. Option B is incorrect because Groovy rules are not created by data scientists to evaluate historical data; instead, these rules are automatically generated by the Planning application during the import process to integrate the ML model with the application. Option C is also incorrect, as EPM administrators do not generate PMML files by creating data models and pushing data—instead, they import an existing PMML file. Finally, Option D is incorrect because while EPM administrators may create data maps and Groovy rules as part of the deployment process, this occurs after the PMML file is imported, not before.

The Oracle Planning 2024 Implementation documentation emphasizes that the "Bring Your Own ML" functionality relies on importing a prebuilt PMML file, making the data scientists' role in building and training the model a mandatory first step.