In order to develop suicide prevention systems in the network, a pilot corpus of suicide thoughts was compiled and annotated. It was extracted from social networks. Texts has been obtained both from the Web and Deep Web. The selected written texts are in Spanish and English. Therefore, to characterize semantically each message, these have been annotated according to suicide relationship. The corpus compilation process ensures the representativeness of the texts and the consistent annotation between annotations.

Source Type: Journal

Original language: Spanish

Document Type: Article

Publisher: Sociedad Española para el Procesamiento del Lenguaje Natural