A manufacturing company uses an ML model to determine whether products meet a standard for quality. The model produces an output of "Passed" or "Failed." Robots separate the products into the two…

Question

Options

• APrecision and recall
• BRoot mean square error (RMSE) and mean absolute percentage error (MAPE)
• CAccuracy and F1 score
• DBilingual Evaluation Understudy (BLEU) score
• EPerplexity

• A
  87% (40)
• B
  4% (2)
• D
  7% (3)
• E
  2% (1)

Explanation

This is a binary classification problem - the model outputs one of two discrete labels (Passed/Failed), not a continuous number.

Why A and C are correct: Precision, recall, accuracy, and F1 score are all standard metrics for evaluating classification models. Precision answers "of all items the model flagged as Passed, how many actually passed?" while recall answers "of all actual passing products, how many did the model catch?" Accuracy gives the overall percentage of correct predictions, and F1 score is the harmonic mean of precision and recall - especially useful if the cost of false positives and false negatives differs (e.g., a failed part slipping through vs. a good part being discarded).

Why the distractors are wrong:

• B (RMSE and MAPE): Regression metrics designed for continuous numerical outputs (e.g., predicting temperature or price), not discrete labels.
• D (BLEU score): Used to evaluate the quality of machine-generated text translations, comparing output to a reference translation - completely unrelated to image classification.
• E (Perplexity): A metric for language models, measuring how well a model predicts a sequence of words - also unrelated here.

Memory tip: If your model output is a label or category, think Classification metrics (Precision, Recall, Accuracy, F1). If it's a number, think Regression metrics (RMSE, MAPE). If it's text output, think NLP metrics (BLEU, Perplexity).

No community discussion yet for this question.