3

All other algorithm scores have names that make sense, but the F-score is just "F". Was the letter chosen at random?

gung - Reinstate Monica
  • 132,789
  • 81
  • 357
  • 650
T. Spikes
  • 57
  • 1
  • 4
  • 5
    'all other algorithm scores have names that make sense' what sense would you give to the 'z'-score ? the 't'-score ? –  Sep 01 '17 at 16:48
  • 4
    It would be helpful if you made explicit in your question that you are asking for the F1 score, the harmonic mean of precision and recall, as people are answering about something completely different. – mdewey Sep 01 '17 at 17:06
  • I think you mean F-measure and not score – Cybercop Sep 01 '17 at 19:50
  • @Cybercop I think both names are in widespread use. For what it's worth, the wikipedia article is called "[F1 score](https://en.wikipedia.org/wiki/F1_score)" but lists "F-score" and "F-measure" as alternative names. Obviously we can't do a google n-gram search to establish which is more common, since the result is contaminated by the other type of [F-score](https://en.wikipedia.org/wiki/F-test), which makes this question text rather ambiguous until you look at the tags that have been used! – Silverfish Sep 01 '17 at 22:48
  • 3
    -1 for being unclear exactly what "F-score" is being asked about, not even providing any context clues. – Jake Westfall Sep 02 '17 at 01:11

2 Answers2

12

When F-scores are used on tests and the null hypothesis is true, F-scores (usually) follow an F-distribution, which is said to be named after Ronald Fisher, who developed its main applications - notably ANOVA.

Pere
  • 5,875
  • 1
  • 13
  • 29
  • 17
    I have a feeling that the OP is referring to [this F-score](https://en.wikipedia.org/wiki/F1_score), not [that F-score](https://en.wikipedia.org/wiki/F-test). Unfortunately the only clue for that is in the tags that have been used, not the question text... – Silverfish Sep 01 '17 at 19:05
  • 1
    @Silverfish You may be right. Another answer on that F-score would be interesting, too. – Pere Sep 01 '17 at 22:16
3

If the "F-score" you're referring to is this one then according to these lecture notes the answer appears to be that it is an accident of history.

There is one thing that remains unsolved, which is why the F-measure is called F. A personal communication with David D. Lewis several years ago revealed that when the F-measure was introduced to MUC-4, the name was accidentally selected by the consequence of regarding a different F function in van Rijsbergen’s book as the definition of the “F-measure”.

A Person
  • 330
  • 3
  • 8