An ML engineer is training an ML model to identify people's health risk based on 20 features and 1 target. The target class has two values: - Likely to have health risk (positive class) - Unlikely…

Question

Options

• AOversample the positive class for the age range of 40 to 45.
• BUndersample the positive class for the age range of 40 to 45.
• CUndersample the positive class for all age ranges except 40 to 45.
• DOversample the negative class for all age ranges except 40 to 45.

• A
  11% (4)
• B
  81% (30)
• C
  3% (1)
• D
  5% (2)

Explanation

Difference in Proportions of Labels (DPL) of +0.9 for the 40–45 age range means that this subgroup is dramatically over-represented in the positive class compared to all other age ranges-a strong positive bias. To correct this imbalance, the positive class samples specifically within the 40–45 age range must be reduced, which is undersampling. Option A (oversample positive class for 40–45) would worsen the imbalance by adding even more positive samples to an already over-represented group. Option C (undersample positive class for all other ranges) would make the 40–45 overrepresentation even more extreme by shrinking every other group. Option D (oversample the negative class elsewhere) is less targeted and would inflate the dataset size unnecessarily. Undersampling the over-represented subgroup (B) is the most direct correction.

No community discussion yet for this question.