A company is training an image classification model on Amazon SageMaker AI to differentiate between images of cats and dogs. The company is using images from a third-party source. The company needs to

Question

Options

• AUse SageMaker Clarify to check for class imbalance (CI). If the value is close to 1, then the
• BUse SageMaker Clarify to check for class imbalance (CI). If the value is close to 0, then the
• CUse SageMaker Data Wrangler to check for difference in proportions of labels (DPL). If the value
• DUse SageMaker Data Wrangler to check for difference in proportions of labels (DPL). If the value

• B
  88% (21)
• C
  8% (2)
• D
  4% (1)

Explanation

Amazon SageMaker Clarify computes a Class Imbalance (CI) metric that ranges from -1 to +1. A CI value of 0 indicates perfect balance between the two classes (equal numbers of cat and dog images). A value close to +1 or -1 indicates severe imbalance (one class dominates). Therefore, the correct approach is: run Clarify, check CI, and if CI is close to 0 the dataset is balanced and no further labeling is needed; if it deviates from 0, more images of the underrepresented class must be added. Option A incorrectly states that CI close to 1 means balance - that actually signals imbalance. Options C and D describe SageMaker Data Wrangler's DPL metric, which measures label proportion differences in supervised datasets and is not the right tool for this binary class balance check.

No community discussion yet for this question.