[Corpora-List] Comprehensive Multiword Expressions (CMWE) Corpus

[Nathan Schneider]

Greetings,

The Comprehensive Multiword Expressions (CMWE) Corpus consists of 55,000
words of English web reviews that have been manually annotated for
heterogeneous multiword expressions.

Annotations are shallow but comprehensive: proceeding sentence by sentence,
our annotators grouped tokens into MWEs according to guidelines that cover
a broad range of multiword phenomena—including (but not limited to)
compound nominals, light verb constructions, verb-particle constructions,
prepositional verbs, and multiword named entities. 3,500 MWE instances are
marked, 500 of which are discontinuous (contain a gap). The annotation
scheme makes a qualitative distinction between "strong" (highly
idiosyncratic) and "weak" (loosely collocational) expressions.

For example,

    I will sum_ it _up~with , it was worth_every_penny !

is annotated as containing 2 strong MWEs (sum_up, worth_every_penny) and 1
weak MWE (sum_up~with). Every sentence was reviewed by at least two
annotators.

This resource is described in the paper:

Nathan Schneider, Spencer Onuffer, Nora Kazour, Emily Danchik, Michael T.
Mordowanec, Henrietta Conrad, and Noah A. Smith. Comprehensive annotation
of multiword expressions in a social web
corpus<http://www.cs.cmu.edu/%7Enschneid/mwecorpus.pdf>
. LREC 2014.

The annotations can be downloaded from:

    http://www.ark.cs.cmu.edu/LexSem/

That page also links to annotation guidelines, as well as an open source
MWE identification tool trained and evaluated on the corpus.

Cheers,
Nathan & collaborators at CMU
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listserv.linguistlist.org/pipermail/corpora/attachments/20140528/c984a400/attachment.htm>
-------------- next part --------------
_______________________________________________
UNSUBSCRIBE from this page: http://mailman.uib.no/options/corpora
Corpora mailing list
Corpora at uib.no
http://mailman.uib.no/listinfo/corpora