Using Unlabeled MEDLINE Abstracts for Biological Named Entity Classification
Manabu Torii, K. Vijay‐Shanker · 2002
Named Entity Recognition is a crucial step for Information Extraction from biological texts. By using surface clues such as capitalization, numbers, and special symbols, existing tools extract names of protein and other biological entities well. However, names of different entities share surface characteristics, and it is difficult to classify detected names based only on that attribute. As pointed out by