| Título |
Binary and Multi-class Classification of Lexical Functions in Spanish Verb-Noun Collocations |
| Tipo |
Congreso |
| Sub-tipo |
Indefinido |
| Descripción |
16th Mexican International Conference on Artificial Intelligence, MICAI 2017 |
| Resumen |
Collocations as semi-fixed lexical combinations present a challenge in natural language processing. While collocation identification on the shallow level is a task in which a significant advance has been reached, a deeper semantic representation and analysis of collocations remains an open issue. One of the possible solutions is detection of lexical functions of the Meaning-Text Theory in collocations thus resolving their semantic interpretation. We experimented with four lexical functions (Oper1, Real1, CausFunc0, and CausFunc1) for the special case of Spanish verb-noun collocations. In our experiments we also identified free verb-noun combinations as opposed to lexical functions. We used WordNet hypernyms as features and various algorithms of supervised machine learning; the best result with an F-measure of 0.873 was achieved for detecting Oper1 in binary classification. |
| Observaciones |
DOI 10.1007/978-3-030-02840-4_1
Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), v. 10633 |
| Lugar |
Ensenada |
| País |
Mexico |
| No. de páginas |
3-14 |
| Vol. / Cap. |
10633 LNAI |
| Inicio |
2017-10-23 |
| Fin |
2017-10-28 |
| ISBN/ISSN |
9783030028398 |
