Resumen

KOLESNIKOVA, Olga; GELBUKH, Alexander  y  CHANONA-HERNANDEZ, Liliana. Automatic Detection of Semantic Classes of Verb-Noun Collocations. Comp. y Sist. [online]. 2020, vol.24, n.1, pp.141-150.  Epub 27-Sep-2021. ISSN 2007-9737.  https://doi.org/10.13053/cys-24-1-3130.

It does not surprise us that a bank can be a financial institution as well as a piece of land. Quite often one word is used with different meanings. But sometimes the opposite happens: we choose different words to express the same idea. For example, to give a smile means 'to smile', and to lend support means 'to support' (Longman Dictionary of Contemporary English, 1995). These two collocations convey the same idea: to smile is to 'perform', or 'do' a smile, and to support is to 'do' support, so that both verb-noun collocations share the same semantics: to do what is denoted by the noun. Likewise, we find that to acquire popularity and to sink into despair both mean 'to begin to experience the <noun>', and to establish a relation and to find a solution mean 'to create the <noun>'. Such semantic patterns or classes are called lexical functions. In this article, we explain the concept of lexical functions, give a summary of state-of-the-art research on automatic detection of lexical functions, and present the framework and results of our experiments on supervised learning of lexical functions fulfilled on the material of Spanish verb-noun collocations.

Palabras llave : Verb-noun collocations; lexical functions; semantic classification; supervised machine learning.