Databricks Machine Learning Professional Question 131
Single answerA data science team has built a machine learning model to predict customer churn for a subscription-based service. They want to use these predictions to send targeted retention emails to customers who are likely to churn. The team plans to generate predictions for all customers at the start of each month and update their email campaign accordingly. Which deployment approach is most appropriate for this use case?
- A
Real-time deployment using a REST API endpoint
- B
Batch deployment on a schedule
- C
Streaming deployment for real-time customer activity
- D
Edge deployment on customer devices
Show answer and explanation
Correct answer: B
Explanation
Batch deployment is the appropriate choice for use cases where predictions are needed periodically or in bulk. In this scenario, predictions are required at the start of each month for all customers, making batch deployment the most efficient and practical approach. Real-time, streaming, or edge deployment methods would introduce unnecessary complexity without addressing the specific requirements of the use case.
- A. Incorrect.
Real-time deployment using a REST API endpoint would be unnecessary in this use case since predictions only need to be generated once a month, not in real-time.
- B. Correct.
Batch deployment on a schedule is the most appropriate approach since the predictions are generated periodically (monthly) and do not require real-time processing.
- C. Incorrect.
Streaming deployment for real-time customer activity is not suitable because the use case does not involve continuous incoming data or real-time decision-making.
- D. Incorrect.
Edge deployment on customer devices is irrelevant here as predictions are generated centrally and there is no need to deploy the model on individual devices.