An ML engineer at an insurance company trains a regression model to predict the number of insurance policy sales each month. After training the model, the ML engineer uses Amazon SageMaker AI to…

Question

Options

• ADetermine whether there is drift in the data quality.
• BDetermine whether there is drift in the model quality.
• CDetermine whether there is drift in the model bias.
• DDetermine whether there is drift in the feature attribution.

• A
  79% (23)
• B
  10% (3)
• C
  7% (2)
• D
  3% (1)

Explanation

The scenario describes a shift in the statistical distribution of incoming production data (input features) relative to the training data, caused by changes in customer behavior. In Amazon SageMaker Model Monitor, this is called data quality drift - it monitors statistical properties (mean, standard deviation, distribution) of input features over time and alerts when they deviate from a baseline established at training time. Model quality drift (B) measures prediction accuracy against ground-truth labels, which are not mentioned here. Model bias drift (C) tracks fairness metrics across demographic groups. Feature attribution drift (D) monitors SHAP value changes. Since the concern is that the raw input data distribution is changing - not that labels or bias are shifting - data quality drift monitoring is the correct choice.

No community discussion yet for this question.