A machine learning (ML) specialist wants to create a data preparation job that uses a PySpark script with complex window aggregation operations to create data for training and testing. The ML…

Question

Options

• AAdd an Amazon SageMaker Debugger hook to the script to capture key metrics. Run the
• BAdd an Amazon SageMaker Experiments tracker to the script to capture key metrics. Run
• CAdd an Amazon SageMaker Debugger hook to the script to capture key parameters. Run
• DAdd an Amazon SageMaker Experiments tracker to the script to capture key parameters.

• A
  3% (2)
• B
  80% (48)
• C
  5% (3)
• D
  12% (7)

Explanation

Amazon SageMaker Experiments (Option B) is the correct service for this task. It provides a structured framework to log, compare, and visualize the parameters and metrics across multiple experiment runs - exactly what is needed to evaluate how varying the number of features and sample counts affects model performance. Metrics (accuracy, loss, etc.) captured during each run are stored and searchable, enabling systematic comparison. Amazon SageMaker Debugger (Options A and C) is designed to detect training anomalies like vanishing gradients or overfitting during a running training job - it is a debugging tool, not an experiment tracking tool. Option D describes SageMaker Experiments but focuses on 'parameters' only; Option B correctly captures both metrics and parameters, making it the more complete and accurate choice for evaluating the impact on model performance.

No community discussion yet for this question.