F2-Measure

From GM-RKB
(Redirected from F2 Metric)
Jump to navigation Jump to search

An F2-Measure is an F-measure where the β is equal to 2.

  • AKA: [math]\displaystyle{ F_2 }[/math].
  • Context:
    • It can be calculated as:
      • [math]\displaystyle{ F_2 }[/math] = (1 + 2²) / ((1/P) + (2²/R)).
      • [math]\displaystyle{ F_2 }[/math] = 5PR / (4P + R).
      • [math]\displaystyle{ F_2 }[/math] = 5TP / (5TP + 4FP + FN).
    • It can (typically) place more emphasis on R than P, making it more useful in situations where missing positive cases (high FN) is more costly.
    • It can (often) be used in scenarios like medical diagnosis or document retrieval where failing to identify a relevant case (low recall) is more detrimental than incorrectly identifying irrelevant cases (lower precision).
    • ...
  • Counter-Example(s):
  • See: Harmonic Mean, Precision, Recall, True Positives, False Positives, False Negatives.


References