MLA-C01 Question 70
Select 2You are building a real-time machine learning pipeline to analyze customer sentiment from a stream of social media posts. The pipeline must first preprocess the incoming data, such as removing special characters, and then apply a trained machine learning model for sentiment analysis. Which combination of AWS services can you use to meet these requirements?
- A
AWS Lambda for data preprocessing and Amazon SageMaker Endpoint for inference
- B
Apache Spark on Amazon EMR for data preprocessing and Amazon Rekognition for inference
- C
AWS Glue ETL for data preprocessing and AWS Lambda for inference
- D
Kinesis Data Analytics for data preprocessing and Amazon SageMaker Endpoint for inference
- E
AWS Batch for data preprocessing and Kinesis Data Streams for inference
Show answer and explanation
Correct answers: A, D
Explanation
To transform streaming data for real-time machine learning, you need services that can handle both preprocessing and inference efficiently. AWS Lambda and Kinesis Data Analytics are suitable for real-time streaming data preprocessing, while Amazon SageMaker Endpoints are optimized for hosting trained machine learning models for inference. This combination ensures low-latency processing and prediction, which is critical for real-time applications like sentiment analysis.
- A. Correct.
Correct: AWS Lambda is well-suited for lightweight, event-driven data preprocessing tasks, and Amazon SageMaker Endpoints are ideal for hosting trained models to perform real-time inference.
- B. Incorrect.
Incorrect: While Apache Spark on Amazon EMR is powerful for batch processing or large-scale distributed computing, it is not typically used for real-time streaming. Amazon Rekognition is intended for image and video analysis, not text sentiment analysis.
- C. Incorrect.
Incorrect: AWS Glue ETL is designed for batch-oriented extract-transform-load workflows, not real-time streaming. AWS Lambda, while capable of inference, is not optimized for hosting trained ML models for high-frequency real-time predictions.
- D. Correct.
Correct: Kinesis Data Analytics can process and transform streaming data in real time, making it suitable for preprocessing tasks. Amazon SageMaker Endpoints can then be used for real-time inference.
- E. Incorrect.
Incorrect: AWS Batch is meant for running batch computing jobs, which is unsuitable for real-time data processing. Kinesis Data Streams is used for ingesting and streaming data but does not perform inference.