Event Annotation Schemes and Event Recognition in Spanish Texts

• Published in: Computational Linguistics and Intelligent Text Processing pp. 206 - 218
• Main Authors: Wonsever, Dina, Rosá, Aiala, Malcuori, Marisa, Moncecchi, Guillermo, Descoins, Alan
• Format: Book Chapter
• Language: English
• Published: Berlin, Heidelberg Springer Berlin Heidelberg 2012
• Series: Lecture Notes in Computer Science
• Subjects: Conditional Random Field; Factivity Attribute; Natural Language Processing; Support Vector Machine; Training Corpus
• ISBN: 9783642286001; 3642286003
• ISSN: 0302-9743; 1611-3349
• DOI: 10.1007/978-3-642-28601-8_18

This paper presents an annotation scheme for events in Spanish texts, based on TimeML for English. This scheme is contrasted with different proposals, all of them based on TimeML, for various Romance languages: Italian, French and Spanish. Two manually annotated corpora for Spanish, under the proposed scheme, are now available. While manual annotation is far from trivial, we obtained a very good event identification agreement (93% of events were identically identified by both annotators). Part of the annotated text was used as a training corpus for the automatic recognition of events. In the experiments conducted so far (SVM and CRF) our best results are in the state of the art for this task (80.3% of F-measure).