Which combination of steps should the company take to use Amazon Redshift ML to meet these requirements?
(Choose three.)
Define the feature variables and target variable for the churn prediction model.
Use the SOL EXPLAIN_MODEL function to run predictions.
Write a CREATE MODEL SQL statement to create a model.
Use Amazon Redshift Spectrum to train the model.
Manually export the training data to Amazon S3.
Use the SQL prediction function to run predictions.
Explanations:
Defining the feature variables and target variable is a crucial step in any machine learning model development process. This step involves identifying which columns in the dataset will serve as inputs (features) and which will be the output (target) that the model aims to predict—in this case, customer churn.
The SQL EXPLAIN_MODEL function is not used for running predictions. Instead, it is used to analyze and understand the model’s structure and performance. Predictions in Amazon Redshift ML are made using specific SQL prediction functions, not with the EXPLAIN_MODEL function.
The CREATE MODEL SQL statement is used to create a machine learning model in Amazon Redshift ML. This statement incorporates the training data, feature variables, and target variable to train the model within the Redshift environment.
Amazon Redshift Spectrum allows querying data stored in S3, but it is not necessary to use it specifically to train a model in Amazon Redshift ML. Redshift ML can train models directly using the data already present in the Redshift data warehouse without needing to utilize Spectrum for this purpose.
Manually exporting training data to Amazon S3 is not required for using Amazon Redshift ML, as the model can be trained directly on data within the Redshift data warehouse. Redshift ML is designed to work with data stored in Redshift without the need for exporting to S3.
The SQL prediction function in Amazon Redshift ML is used to run predictions based on the model created. This function allows users to input new data and get predictions regarding customer churn directly within the data warehouse environment.