DMCA
TR-NLP-UT-2006-1 * GENIA Corpus Manual Encoding schemes for the corpus and annotation (2006)
| Citations: | 1 - 0 self |
BibTeX
@MISC{Kim06tr-nlp-ut-2006-1*,
author = {Jin-dong Kim and Tomoko Ohta and Yuka Teteisi and Genia Corpus},
title = {TR-NLP-UT-2006-1 * GENIA Corpus Manual Encoding schemes for the corpus and annotation},
year = {2006}
}
Share
 |
 |
 |
 |
||
OpenURL
Abstract
The GENIA corpus a product of the GENIA project of which the objective is to develop information extraction (IE) and text mining (TM) systems for the specific subject domain of molecular biology and medical science. The purpose of developing the corpus is to provide reference materials for training and evaluation of the IE and TM systems. 1.1. Compilation Currently, the GENIA corpus is made up of the titles and abstracts of journal articles which have been taken from the Medline database. Whereas the Medline indexes broad range of academic articles covering the general or specific domains of life sciences, GENIA is intended to cover a smaller subject domain: biological reactions concerning transcription factors in human blood cells. To retrieve the corresponding records from Medline, we used the following search query on the PubMed web interface: “Humans”[MeSH] AND “Blood Cells”[MeSH] AND “Transcription Factors”[MeSH]. 1.2. Annotation The GENIA corpus has been being annotated by encoding human’s interpretation of the text. The purpose of the annotation is two fold. First, we annotate the corpus to make the biomedical knowledge encoded in the text transparent. Second, we also annotate the corpus to reveal the syntactic structure behind the text. Eventually, our objective is to establish the mapping between the knowledge pieces and the linguistic structures. 2 2. Document Encoding Currently, the source of the GENIA corpus is the Medline database. Each Medline record can be retrieved from PubMed by using its PMID (PubMed ID). Accordingly, a Medline record may have its URI like as follows:
Keyphrases
genia corpus medline record medline database human blood cell life science pubmed web interface syntactic structure text mining transcription factor compilation currently corresponding record molecular biology journal article specific domain knowledge piece following search query human interpretation reference material biological reaction pubmed id document encoding currently medical science tm system subject domain text transparent medline index broad range information extraction specific subject domain linguistic structure genia project academic article biomedical knowledge
