So far I understood, that hyphenation should aid readability.

Examples [1, 2]:

North America-based company

A Gaussian mixture model-based approach

We propose spherical Gaussian-based approximations to calculate this analytically.

Although, this never aligned with my understanding of parsing trees, I would still like to apply this rule.

How does it extend to abbreviation remarks?

Gaussian mixture model (GMM)-based approach

Non-negative matrix factorization (NMF)-inspired method

My own understanding of how to parse the words is as follows, which does not seem to be reflected in how hyphens are used:

      Gaussian {
        mixture model
    } (GMM)
} approach

Hyphens are used to compose constituents, either words or phrases, to make words. Consequently, to know whether a hyphen is appropriate, you have to know the categories of constituents, not just what the constituents are. Below, I've tried to amend your diagram for "Gaussian mixture model (GMM)-based approach" by adding category (parts of speech) information. NP means noun phrase, N is noun (a word), A is adjective or other noun-modifier (a word), Participle (a word).

      A Gaussian {N
        N mixture N model
    } (GMM)
  }-Participle based
} N approach  

There are two types of word compounds in the example. A compound adjective (a word) is made by combining a NP (a phrase) and a Participle (a word), and a compound N (a word) is made by combining two Ns (words). For the latter type of compound, a hyphen is often optional.

I'm not sure I see a problem with the hyphenation. I'm worried, though, about the structure of "Gaussian mixture model", which must be a phrase, not a single word, because "Gaussian" is an adjective, and noun-noun compounds can't contain adjectives. But "Gaussian mixture" should be a constituent, because of the interpretation: mixture of Gaussian distributions.

| improve this answer | |
  • The first 3 example sentences in your question looked spot on to me. The later versions looked ugly. // I might have written: 'GMM' or Gaussian mixture model-based approach. After defining my abbreviation with quote marks the first time it was used I would later use simply: GMM-based. – Ross Murray Feb 1 '18 at 14:14
  • @RossMurray, I guess this comment on my answer is intended to be a comment on the question, instead. – Greg Lee Feb 1 '18 at 16:20
  • YES, and I apologise. // I am new enough here that this is the first time I have made that particular mistake . // Can that be fixed by the moderators? // I do not know how to gain their attention. I wiped out somebody else's answer the last time I tried. :( // You have my blessing to ask them. – Ross Murray Feb 1 '18 at 20:29
  • @RossMurray, No matter. These comments have made clear what happened. – Greg Lee Feb 1 '18 at 20:35

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.