MEDLINE MeSH Indexing: Lessons Learned from Machine Learning and Future Directions

Antonio Jimeno-Yepes; James G. Mork; Bartlomiej Wilkowski; Dina Demner Fushman; Alan R. Aronson

doi:10.1145/2110363.2110450

MEDLINE MeSH Indexing: Lessons Learned from Machine Learning and Future Directions

Antonio Jimeno-Yepes, James G. Mork, Bartlomiej Wilkowski, Dina Demner Fushman, Alan R. Aronson

U.S National Library of Medicine

Research output: Chapter in Book/Report/Conference proceeding › Article in proceedings › Research › peer-review

Abstract

Due to the large yearly growth of MEDLINE, MeSH indexing is becoming a more difficult task for a relatively small group of highly qualified indexing staff at the US National Library of Medicine (NLM). The Medical Text Indexer (MTI) is a support tool for assisting indexers; this tool relies on MetaMap and a k-NN approach called PubMed Related Citations (PRC). Our motivation is to improve the quality of MTI based on machine learning. Typical machine learning approaches fit this indexing task into text categorization. In this work, we have studied some Medical Subject Headings (MeSH) recommended by MTI and analyzed the issues when using standard machine learning algorithms. We show that in some cases machine learning can improve the annotations already recommended by MTI, that machine learning based on low variance methods achieves better performance and that each MeSH heading presents a different behavior. In addition, there are several factors which make this task difficult (e.g. limited access to the full-text of the citations) which provide direction for future work.

Original language	English
Title of host publication	IHI '12 Proceedings of the 2nd ACM SIGHIT International Health Informatics Symposium
Publisher	Association for Computing Machinery
Publication date	2012
Pages	737-742
ISBN (Print)	978-1-4503-0781-9
DOIs	https://doi.org/10.1145/2110363.2110450
Publication status	Published - 2012
Event	2nd ACM SIGHIT International Health Informatics Symposium (IHI 2012) - Miami, Florida, United States Duration: 28 Jan 2012 → 30 Jan 2012 http://www.sighit.org/ihi2012/

Conference

Conference	2nd ACM SIGHIT International Health Informatics Symposium (IHI 2012)
Country/Territory	United States
City	Miami, Florida
Period	28/01/2012 → 30/01/2012
Internet address	http://www.sighit.org/ihi2012/

Keywords

Database systems
Indexing (of information)
Learning systems
Text processing
Learning algorithms

Access to Document

10.1145/2110363.2110450

OpenUrl availability

Full text

Cite this

@inproceedings{5c17d99d8bfc4d6e8e658f0c737252b2,

title = "MEDLINE MeSH Indexing: Lessons Learned from Machine Learning and Future Directions",

abstract = "Due to the large yearly growth of MEDLINE, MeSH indexing is becoming a more difficult task for a relatively small group of highly qualified indexing staff at the US National Library of Medicine (NLM). The Medical Text Indexer (MTI) is a support tool for assisting indexers; this tool relies on MetaMap and a k-NN approach called PubMed Related Citations (PRC). Our motivation is to improve the quality of MTI based on machine learning. Typical machine learning approaches fit this indexing task into text categorization. In this work, we have studied some Medical Subject Headings (MeSH) recommended by MTI and analyzed the issues when using standard machine learning algorithms. We show that in some cases machine learning can improve the annotations already recommended by MTI, that machine learning based on low variance methods achieves better performance and that each MeSH heading presents a different behavior. In addition, there are several factors which make this task difficult (e.g. limited access to the full-text of the citations) which provide direction for future work.",

keywords = "Database systems, Indexing (of information), Learning systems, Text processing, Learning algorithms",

author = "Antonio Jimeno-Yepes and Mork, {James G.} and Bartlomiej Wilkowski and Fushman, {Dina Demner} and Aronson, {Alan R.}",

year = "2012",

doi = "10.1145/2110363.2110450",

language = "English",

isbn = "978-1-4503-0781-9",

pages = "737--742",

booktitle = "IHI '12 Proceedings of the 2nd ACM SIGHIT International Health Informatics Symposium",

publisher = "Association for Computing Machinery",

address = "United States",

note = "2nd ACM SIGHIT International Health Informatics Symposium (IHI 2012) ; Conference date: 28-01-2012 Through 30-01-2012",

url = "http://www.sighit.org/ihi2012/",

}

Jimeno-Yepes, A, Mork, JG, Wilkowski, B, Fushman, DD & Aronson, AR 2012, MEDLINE MeSH Indexing: Lessons Learned from Machine Learning and Future Directions. in IHI '12 Proceedings of the 2nd ACM SIGHIT International Health Informatics Symposium. Association for Computing Machinery, pp. 737-742, 2nd ACM SIGHIT International Health Informatics Symposium (IHI 2012), Miami, Florida, United States, 28/01/2012. https://doi.org/10.1145/2110363.2110450

MEDLINE MeSH Indexing: Lessons Learned from Machine Learning and Future Directions. / Jimeno-Yepes, Antonio; Mork, James G.; Wilkowski, Bartlomiej et al.
IHI '12 Proceedings of the 2nd ACM SIGHIT International Health Informatics Symposium. Association for Computing Machinery, 2012. p. 737-742.

Research output: Chapter in Book/Report/Conference proceeding › Article in proceedings › Research › peer-review

TY - GEN

T1 - MEDLINE MeSH Indexing: Lessons Learned from Machine Learning and Future Directions

AU - Jimeno-Yepes, Antonio

AU - Mork, James G.

AU - Wilkowski, Bartlomiej

AU - Fushman, Dina Demner

AU - Aronson, Alan R.

PY - 2012

Y1 - 2012

N2 - Due to the large yearly growth of MEDLINE, MeSH indexing is becoming a more difficult task for a relatively small group of highly qualified indexing staff at the US National Library of Medicine (NLM). The Medical Text Indexer (MTI) is a support tool for assisting indexers; this tool relies on MetaMap and a k-NN approach called PubMed Related Citations (PRC). Our motivation is to improve the quality of MTI based on machine learning. Typical machine learning approaches fit this indexing task into text categorization. In this work, we have studied some Medical Subject Headings (MeSH) recommended by MTI and analyzed the issues when using standard machine learning algorithms. We show that in some cases machine learning can improve the annotations already recommended by MTI, that machine learning based on low variance methods achieves better performance and that each MeSH heading presents a different behavior. In addition, there are several factors which make this task difficult (e.g. limited access to the full-text of the citations) which provide direction for future work.

AB - Due to the large yearly growth of MEDLINE, MeSH indexing is becoming a more difficult task for a relatively small group of highly qualified indexing staff at the US National Library of Medicine (NLM). The Medical Text Indexer (MTI) is a support tool for assisting indexers; this tool relies on MetaMap and a k-NN approach called PubMed Related Citations (PRC). Our motivation is to improve the quality of MTI based on machine learning. Typical machine learning approaches fit this indexing task into text categorization. In this work, we have studied some Medical Subject Headings (MeSH) recommended by MTI and analyzed the issues when using standard machine learning algorithms. We show that in some cases machine learning can improve the annotations already recommended by MTI, that machine learning based on low variance methods achieves better performance and that each MeSH heading presents a different behavior. In addition, there are several factors which make this task difficult (e.g. limited access to the full-text of the citations) which provide direction for future work.

KW - Database systems

KW - Indexing (of information)

KW - Learning systems

KW - Text processing

KW - Learning algorithms

U2 - 10.1145/2110363.2110450

DO - 10.1145/2110363.2110450

M3 - Article in proceedings

SN - 978-1-4503-0781-9

SP - 737

EP - 742

BT - IHI '12 Proceedings of the 2nd ACM SIGHIT International Health Informatics Symposium

PB - Association for Computing Machinery

T2 - 2nd ACM SIGHIT International Health Informatics Symposium (IHI 2012)

Y2 - 28 January 2012 through 30 January 2012

ER -

MEDLINE MeSH Indexing: Lessons Learned from Machine Learning and Future Directions

Abstract

Conference

Keywords

Access to Document

OpenUrl availability

Fingerprint

Cite this