Databricks Machine Learning Professional Question 128
Single answerA retail company uses a machine learning model to predict product demand across its stores for the upcoming week. Predictions are generated once a week and used to optimize inventory planning. Which deployment approach is most appropriate for this use case?
- A
Real-time deployment using REST API
- B
Batch deployment on a weekly schedule
- C
Streaming deployment for continuous predictions
- D
Deploying the model as a mobile application for store managers
Show answer and explanation
Correct answer: B
Explanation
Batch deployment is the most appropriate choice for scenarios where predictions are generated at regular intervals (e.g., weekly) and latency is not a concern. In this case, the retail company generates predictions once a week to optimize inventory planning, which aligns perfectly with a batch deployment approach.
- A. Incorrect.
Real-time deployment using REST API is not the best choice for this use case because predictions are needed only once a week, not instantaneously.
- B. Correct.
Batch deployment on a weekly schedule is the correct answer as it matches the requirement to generate predictions once a week. Batch deployment is suitable for periodic prediction tasks where latency is not critical.
- C. Incorrect.
Streaming deployment for continuous predictions is unnecessary here because the prediction schedule is weekly, not real-time or continuous.
- D. Incorrect.
Deploying the model as a mobile application for store managers is not relevant to the given scenario, as the task requires centralized weekly predictions rather than a store-level tool.