Naslov
Determining the Semantic Compositionality of Croatian Multi-Word Expressions

Autori
Almić, Petra ; Šnajder, Jan

Vrsta, podvrsta i kategorija rada
Radovi u zbornicima skupova, cjeloviti rad (in extenso), znanstveni

Izvornik
Proceedings of the Ninth Language Technologies Conference, Information Society (IS-JT 2014) / - Ljubljana, 2014, 32-37

Skup
Ninth Language Technologies Conference, Information Society (IS-JT 2014)

Mjesto i datum
Ljubljana, Slovenija, 09.10.2014. - 10.10.2014

Vrsta sudjelovanja
Predavanje

Vrsta recenzije
Međunarodna recenzija

Ključne riječi
multiword expressions; semantic compositionality; distributional semantics; Croatian language

Sažetak
A distinguishing feature of many multiword expressions (MWEs) is their semantic non-compositionality. Being able to automatically determine the semantic (non-)compositionality of MWEs is important for many natural language processing tasks. We address the task of determining the semantic compositionality of Croatian MWEs. We adopt a composition-based approach within the distributional semantics framework. We build a small dataset of Croatian MWE with human-annotated semantic compositionality scores. We build and evaluate a model for predicting the semantic compositionality based on Latent Semantic Analysis. The predicted scores correlate well with human judgments (rho=0.48). When compositionality detection is treated as a classification task, the model achieves an F1-score of 0.65.

Izvorni jezik
Engleski

Znanstvena područja
Računarstvo

POVEZANOST RADA