The maximum value for the training data occurs for model number 46, and the maximum value for the validation data occurs for model number 43. If you base model selection solely on overall average prof

Question

Options

• ASelect model 46
• BSelect model 43
• CSelect model 45
• DSelect model 21

• A
  14% (7)
• B
  74% (37)
• C
  4% (2)
• D
  8% (4)

Explanation

Selecting model 43 is correct because the question asks for the best overall average profit-meaning the combined average of both training and validation performance. Even though model 46 peaks on training data alone, model 43's superior validation score gives it the higher combined average across both datasets, making it the best choice when both are weighted equally.

Why the distractors fail:

• A (Model 46) is the classic overfitting trap - it wins on training data only, which suggests it may be too tightly fit to that data. Optimizing solely for training performance ignores how the model generalizes.
• C (Model 45) is a plausible-sounding middle ground, but it doesn't lead on either metric, so it can't win the average either.
• D (Model 21) is a far-off distractor with no supporting evidence from either metric.

Memory tip: Think of it as a two-event competition - training and validation are each one event. The winner isn't whoever dominates one event (model 46), but whoever has the best combined score across both. Model 43 is the overall athlete, not just a training-day specialist.

No community discussion yet for this question.