In this paper we discuss five different corpora annotated for protein names. We present several within- and cross-dataset protein tagging experiments showing that different annotation schemes severely affect the portability of statistical protein taggers. By means of a detailed error analysis we identify crucial annotation issues that future annotation projects should take into careful consideration.
The Impact of Annotation on the Performance of Protein Tagging in Biomedical Text / Alex B.; Nissim M.; Grover C.. - STAMPA. - (2006). (Intervento presentato al convegno 5th Language Resources and Evaluation Conference tenutosi a Genova nel 22-28 May 2006).
The Impact of Annotation on the Performance of Protein Tagging in Biomedical Text
NISSIM, MALVINA;
2006
Abstract
In this paper we discuss five different corpora annotated for protein names. We present several within- and cross-dataset protein tagging experiments showing that different annotation schemes severely affect the portability of statistical protein taggers. By means of a detailed error analysis we identify crucial annotation issues that future annotation projects should take into careful consideration.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
