A team wants to build a churn-prediction model and run predictions directly inside an Amazon Redshift cluster using Redshift ML. Which steps should they take? (Choose three.)

To build a churn prediction model with Redshift ML, the first step is to identify the features (input columns) and the target variable (churn status) from your Redshift data, which is fundamental for any machine learning task. Next, you create the model directly within Redshift using a CREATE MODEL SQL statement. This statement specifies the data, the target, and the ML framework (e.g., AutoGluon, XGBoost) to use. Redshift ML handles the training process automatically, including data preparation and model deployment. Finally, to get predictions on new data, you use the SQL prediction functions provided by Redshift ML, such as PREDICT, directly within your SQL queries. Incorrect options: SOL EXPLAINMODEL is used for model interpretability, not for executing predictions. Redshift Spectrum is for querying data in S3, not for training models. Redshift ML handles training internally. Manually exporting data to S3 is unnecessary; Redshift ML can train directly from data within Redshift tables.