A university wants to predict whether an applicant will enroll so it can target recruitment efforts. The data scientist has academic histories and wants to build applicant profiles for prediction. Which two steps should the data scientist take to build a model that predicts enrollment likelihood? (Choose two.)

The problem requires predicting enrollment likelihood, which is a binary outcome (enroll or not enroll). This makes it a classification problem, necessitating a classification algorithm. Regression algorithms predict continuous values, which is not applicable here. Forecasting algorithms predict future values based on time-series data, which isn't the primary goal. Since the historical data likely lacks explicit "enrolled" or "not enrolled" labels, Amazon SageMaker Ground Truth is ideal for human-in-the-loop data labeling to create the necessary training dataset. K-means is an unsupervised clustering algorithm, not suitable for supervised prediction of predefined classes like "enrolled" and "not enrolled" without prior labeling.