A Semantic Parser for Neuro-Degenerative Disease Knowledge Discovery

Ever increasing size of the biomedical literature makes tapping into implicit knowledge in scientific literature a necessity for knowledge discovery. In this paper, a semantic parser for recognizing semantic roles and named entities in individual sentences of schizophrenia related scientific abstracts is described. The named entity recognizer, CRFNER, outperforms ABNER in biological named entity recognition and achieves 82.5% micro-averaged F1 on clinical psychology/neuroscience named entities. Support vector machine based semantic role labeling system achieves 75.3% microaveraged F1 for semantic role identification and classification on schizophrenia corpus.

• Type: Conference
• Year: 2008
• Where: FLAIRS
• Authors: I. Burak Ozyurt
• Keywords: Artificial Intelligence, FLAIRS 2008, Named Entity Recognition, Semantic Role, Semantic Role Labeling