5

I made an edit to a question and the portion mentioning the n-gram chart was reverted. I was reading the Wikipedia article about n-grams and added the hyphen based on that usage. This article says Ngram is just Google’s branding for its Ngram Viewer tool. Which spelling is correct?

In my mind n-gram should refer to the word phrase, and Ngram should be rephrased to Ngram comparison when referring to a chart showing the comparison of n-grams.

Community
  • 1
jjclarkson
  • 1,143
  • Maybe I should have done it first but, reading http://english.stackexchange.com/questions/11570/to-hyphenate-or-not now. – jjclarkson Aug 01 '12 at 22:27
  • Google spells it both ways. See this page – Luke_0 Aug 01 '12 at 22:35
  • I'm not sure if this is a peeve or what. The term N-gram is relatively new, and as often happens with such terms, once it starts to get more common, people naturally tend to dispense with the capitalisation and hyphenation. It's meaningless to ask which form is "correct", particularly when here on ELU we're mostly concerned with Google's offering. Which fairly obviously is correctly named Ngram, since that's what Google normally choose to call it. – FumbleFingers Aug 01 '12 at 22:48
  • In context in that answer I should think either "N-grams" or "Google Ngram Viewer" would be right, but not "Ngram". – MetaEd Aug 01 '12 at 23:06
  • I usually refer to it as Ngram, as a shorthand for "the output of the Google books Ngram Viewer". – J.R. Aug 02 '12 at 02:56

2 Answers2

4

  • 'n-gram' is what text analysts use to describe sequences of text units (either characters or terms) of length n

  • 'Ngram' is Google's term for its indexed book text search application.

The two spellings are correct for their given uses. Calling Google's app 'n-gram' and calling a table made of n length sequences an 'Ngram' would both be wrong.

Mitch
  • 71,423
0

After approving your spelling correction, I changed "N-grams" back to "Ngrams" because that is the form most often used in english.stackexchange to refer to "Google Ngram Viewer" results. I didn't recall seeing any ELU posts that used forms other than ngram, ngrams, Ngram, or Ngrams to refer to such results, but did a search and found out that google for the unhyphenated terms shows 1800+ results, vs 22 results for "n-grams". I'm not cognizant of any stronger argument, for one form vs the other, than consistency.

James Waldby - jwpat7
  • 66,660