SAS Certified Associate: Applied Statistics for Machine Learning

A data science team is debating whether the goal of a given project is to understand which factors drive an outcome versus to accurately predict future cases. Under the exam's framework, which task specifically distinguishes the machine learning practice of estimating how well a model will perform on new, unseen data?

An analyst finds that as model complexity increases, the model fits the training data increasingly well but performs worse on holdout data. Which concept from the predictive modeling objectives best describes managing this competing behavior?