Segmentador de oraciones para textos en español basado en red neuronal
- Espinosa García, Jacinto
- Nieto Serrano, Amalio F.
- González, José Carlos
- Goñi Menoyo, José Miguel
ISSN: 1135-5948
Argitalpen urtea: 1996
Zenbakia: 19
Orrialdeak: 1-6
Mota: Artikulua
Beste argitalpen batzuk: Procesamiento del lenguaje natural
Laburpena
This work shows a connectionist system used for the segmentation of texts in Spanish into separate sentences. This task has to be carried out in many Natural Language Processing applications. This kind of pre-processing is not conceptually complex, and several techniques producing acceptable results may be applied. However, the task of text segmentation depends heavily on the sources (structure, layout, genre, style). Most of times, this fact imposes some reworking for every new application and type of text. By using neural nets, low level programming is replaced by learning from sets of sentences correctly classified by a specialist.