Identification of Verbal Phraseological Units in Mexican News Stories

Belém Priego Sánchez, David Pinto


Verbal Phraseological Units are phrasesmade up of two or more words in which at least one of thewords is a verb that plays the role of the predicate. Oneof the characteristics of this type of expression is that itsglobal meaning rarely can be deduced from the meaningof its components. The automatic recognition of this typeof linguistic structures is a very important task, since theyare a standard way of expressing a concept or idea. Inthis paper we present the results obtained when differentsupervised machine learning methods are employed fordetermining whether or not a verbal phraseological unitis present in a given story of a newspaper. The experimentshave been carried out using a supervised corpusof news stories (written in Mexican Spanish). Besidethe results obtained in the experiments aforementioned,we provide access to a new lexicon having phrases asentries (instead of single words), in which each entry isassociated to a real value (normalized between zero andone) indicating its probability of being a verbal phraseologicalunit.


Verbal phraseological units, supervised machine learning, lexicon

Full Text: PDF