Kernel-based learning for biomedical relation extraction

Jiexun Li, Zhu Zhang, Xin Li, Hsinchun Chen

Research output: Contribution to journalArticle

47 Citations (Scopus)

Abstract

Relation extraction is the process of scanning text for relationships between named entities. Recently, significant studies have focused on automatically extracting relations from biomedical corpora. Most existing biomedical relation extractors require manual creation of biomedical lexicons or parsing templates based on domain knowledge. In this study, we propose to use kernel-based learning methods to automatically extract biomedical relations from literature text. We develop a framework of kernel-based learning for biomedical relation extraction. In particular, we modified the standard tree kernel function by incorporating a trace kernel to capture richer contextual information. In our experiments on a biomedical corpus, we compare different kernel functions for biomedical relation detection and classification. The experimental results show that a tree kernel outperforms word and sequence kernels for relation detection, our trace-tree kernel outperforms the standard tree kernel, and a composite kernel outperforms individual kernels for relation extraction.

Original languageEnglish (US)
Pages (from-to)756-769
Number of pages14
JournalJournal of the American Society for Information Science and Technology
Volume59
Issue number5
DOIs
StatePublished - Mar 2008

Fingerprint

learning method
learning
experiment
knowledge
Scanning
Composite materials
Kernel
Experiments
literature

ASJC Scopus subject areas

  • Information Systems
  • Library and Information Sciences

Cite this

Kernel-based learning for biomedical relation extraction. / Li, Jiexun; Zhang, Zhu; Li, Xin; Chen, Hsinchun.

In: Journal of the American Society for Information Science and Technology, Vol. 59, No. 5, 03.2008, p. 756-769.

Research output: Contribution to journalArticle

@article{081907adb05e4afb87267ac0fd01b76e,
title = "Kernel-based learning for biomedical relation extraction",
abstract = "Relation extraction is the process of scanning text for relationships between named entities. Recently, significant studies have focused on automatically extracting relations from biomedical corpora. Most existing biomedical relation extractors require manual creation of biomedical lexicons or parsing templates based on domain knowledge. In this study, we propose to use kernel-based learning methods to automatically extract biomedical relations from literature text. We develop a framework of kernel-based learning for biomedical relation extraction. In particular, we modified the standard tree kernel function by incorporating a trace kernel to capture richer contextual information. In our experiments on a biomedical corpus, we compare different kernel functions for biomedical relation detection and classification. The experimental results show that a tree kernel outperforms word and sequence kernels for relation detection, our trace-tree kernel outperforms the standard tree kernel, and a composite kernel outperforms individual kernels for relation extraction.",
author = "Jiexun Li and Zhu Zhang and Xin Li and Hsinchun Chen",
year = "2008",
month = "3",
doi = "10.1002/asi.20791",
language = "English (US)",
volume = "59",
pages = "756--769",
journal = "Journal of the Association for Information Science and Technology",
issn = "2330-1635",
publisher = "John Wiley and Sons Ltd",
number = "5",

}

TY - JOUR

T1 - Kernel-based learning for biomedical relation extraction

AU - Li, Jiexun

AU - Zhang, Zhu

AU - Li, Xin

AU - Chen, Hsinchun

PY - 2008/3

Y1 - 2008/3

N2 - Relation extraction is the process of scanning text for relationships between named entities. Recently, significant studies have focused on automatically extracting relations from biomedical corpora. Most existing biomedical relation extractors require manual creation of biomedical lexicons or parsing templates based on domain knowledge. In this study, we propose to use kernel-based learning methods to automatically extract biomedical relations from literature text. We develop a framework of kernel-based learning for biomedical relation extraction. In particular, we modified the standard tree kernel function by incorporating a trace kernel to capture richer contextual information. In our experiments on a biomedical corpus, we compare different kernel functions for biomedical relation detection and classification. The experimental results show that a tree kernel outperforms word and sequence kernels for relation detection, our trace-tree kernel outperforms the standard tree kernel, and a composite kernel outperforms individual kernels for relation extraction.

AB - Relation extraction is the process of scanning text for relationships between named entities. Recently, significant studies have focused on automatically extracting relations from biomedical corpora. Most existing biomedical relation extractors require manual creation of biomedical lexicons or parsing templates based on domain knowledge. In this study, we propose to use kernel-based learning methods to automatically extract biomedical relations from literature text. We develop a framework of kernel-based learning for biomedical relation extraction. In particular, we modified the standard tree kernel function by incorporating a trace kernel to capture richer contextual information. In our experiments on a biomedical corpus, we compare different kernel functions for biomedical relation detection and classification. The experimental results show that a tree kernel outperforms word and sequence kernels for relation detection, our trace-tree kernel outperforms the standard tree kernel, and a composite kernel outperforms individual kernels for relation extraction.

UR - http://www.scopus.com/inward/record.url?scp=41449115531&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=41449115531&partnerID=8YFLogxK

U2 - 10.1002/asi.20791

DO - 10.1002/asi.20791

M3 - Article

AN - SCOPUS:41449115531

VL - 59

SP - 756

EP - 769

JO - Journal of the Association for Information Science and Technology

JF - Journal of the Association for Information Science and Technology

SN - 2330-1635

IS - 5

ER -