MLS-C01 Question 24
Select 2A machine learning team is tasked with building a recommendation engine for an e-commerce platform. The platform receives continuous user activity data, such as clicks, searches, and purchases, in real-time. The team also periodically updates the recommendation model using a dataset of user activity aggregated over the past month. Which combination of data job types should the team use to implement this solution effectively?
- A
Batch processing for the periodic model updates
- B
Streaming processing for real-time user activity ingestion
- C
Batch processing for real-time user activity ingestion
- D
Streaming processing for the periodic model updates
- E
Batch processing for one-time historical data analysis
Show answer and explanation
Correct answers: A, B
Explanation
The solution requires two complementary data job styles. Streaming processing is used for real-time ingestion of user activity data, enabling the system to process continuous streams of events with minimal latency. Batch processing is used for periodic model updates, where large datasets accumulated over time are processed to retrain the recommendation model. This combination ensures both real-time responsiveness and periodic model enhancements.
- A. Correct.
Batch processing is well-suited for the periodic model updates since it involves processing large datasets at fixed intervals.
- B. Correct.
Streaming processing is the appropriate choice for real-time user activity ingestion as it handles continuous data flow with low latency.
- C. Incorrect.
Batch processing is not suitable for real-time ingestion as it is designed for processing large datasets in scheduled intervals, not continuous streams.
- D. Incorrect.
Streaming processing is not applicable for periodic model updates because such updates do not require continuous real-time data processing.
- E. Incorrect.
While batch processing can be used for one-time historical data analysis, this scenario does not involve such a task.