AIF-C01 Question 12
Select 1A data science team is developing a machine learning model to predict product demand for a retail company. The team needs to generate predictions for thousands of products daily to plan inventory, and they also require instant predictions when users search for specific products on their website. Which types of inferencing should the team use to meet both requirements?
- A
Batch inferencing for daily predictions and real-time inferencing for user searches
- B
Real-time inferencing for both daily predictions and user searches
- C
Batch inferencing for both daily predictions and user searches
- D
Real-time inferencing for daily predictions and batch inferencing for user searches
Show answer and explanation
Correct answer: A
Explanation
Inferencing in machine learning involves generating predictions from a trained model. Batch inferencing processes large datasets at once and is suitable for scenarios like daily demand forecasting. In contrast, real-time inferencing provides predictions instantly and is appropriate for use cases like user searches on a website. Combining the two types of inferencing allows the team to efficiently meet both the bulk processing and on-demand prediction requirements.
- A. Correct.
This is correct because batch inferencing is suitable for generating large-scale predictions daily, while real-time inferencing is ideal for on-demand predictions such as user searches.
- B. Incorrect.
This is incorrect because real-time inferencing is not efficient for generating bulk predictions like daily demand forecasts, which are better suited for batch processing.
- C. Incorrect.
This is incorrect because batch inferencing cannot handle low-latency, on-demand requests like user searches, making it unsuitable for that use case.
- D. Incorrect.
This is incorrect because real-time inferencing for large-scale daily predictions is inefficient, and batch inferencing cannot handle instant user search predictions.