Combining context and existing knowledge when recognizing biological entities - Early results

Mika Timonen, Antti Pesonen

Research output: Chapter in Book/Report/Conference proceedingChapter or book articleScientificpeer-review

1 Citation (Scopus)

Abstract

Entity recognition has been studied for several years with good results. However, as the focus of information extraction (IE) and entity recognition (ER) has been set on biology and bioinformatics, the existing methods do not produce as good results as before. This is mainly due to the complex naming conventions of biological entities. In our information extraction system for biomedical documents called OAT (Ontology Aided Text mining system) we developed our own method for recognizing the biological entities. The difference to the existing methods, which use lexicons, rules and statistics, is that we combine the context of the entity with the existing knowledge about the relationships of the entities. This has produced encouraging preliminary results. This paper describes the approach we are using in our information extraction system for entity recognition.
Original languageEnglish
Title of host publicationAdvances in Knowledge Discovery and Data Mining
Subtitle of host publication12th Pacific-Asia Conference, PAKDD 2008 Osaka, Japan, May 20-23, 2008 Proceedings
EditorsTakashi Washio, Einoshin Suzuki, Kai Ming Ting, Akihiro Inokuchi
PublisherSpringer
Pages1028-1034
ISBN (Electronic)978-3-540-68125-0
ISBN (Print)978-3-540-68124-3
DOIs
Publication statusPublished - 2008
MoE publication typeA3 Part of a book or another research book
Event12th Pacific-Asia Conference, PAKDD 2008. Osaka, Japan, 20 - 23 May, 2008 -
Duration: 1 Jan 2008 → …

Publication series

SeriesLecture Notes in Computer Science
Volume5102
ISSN0302-9743

Conference

Conference12th Pacific-Asia Conference, PAKDD 2008. Osaka, Japan, 20 - 23 May, 2008
Period1/01/08 → …

Keywords

  • entity recognition
  • entity classification
  • information extraction

Fingerprint

Dive into the research topics of 'Combining context and existing knowledge when recognizing biological entities - Early results'. Together they form a unique fingerprint.

Cite this