5

Does anybody know of source (probably NLP study) of comprehensive list of collocations in English (not only the frequent one). Additionally, does anybody know the approximate estimate of the total number of collocations.

6
3

A great place to start is http://www.collocates.info which is based on COCA (Corpus of Contemporary American English). It is incredibly comprehensive (not free but the prices range from $45-$250).

This site also answers the question as to how many collocates there are in English? 4.3 million. That is from the perspective of all possible collocate relationships (even incredibly weak ones). You can only ask that question meaningfully if you have a purpose in mind. For instance, collocates to help students learn. Collocates that should trigger text autocomplete. Etc.

If you're looking for collocations for students, then you can start with the Oxford Collocations Dictionary which lists 250k collocations.

Or you can use the free online collocations dictionary. But you can only search that, not browse the whole list.

For NLP purposes, your best source is COCA or the already mentioned Google Ngrams (which has many limitations, also mentioned).

1

You can download Google Ngrams whole dataset.

http://storage.googleapis.com/books/ngrams/books/datasetsv2.html

3
  • Unfortunately, n-grams are not collocations, but collocations are n-grams.
    – Ross
    Aug 17 '13 at 7:54
  • I don't know what you mean. Isn't there some way to extract the information you want out of the n-gram dataset?
    – Moss
    Aug 17 '13 at 16:56
  • Collocations are subset of n-grams. I already work with Google Ngrams. That is why I am looking for NLP study that already have extracted collocations.
    – Ross
    Aug 17 '13 at 20:23

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.