Towards constructing a corpus for studying the effects of treatments and substances reported in pubMed abstracts

Evgeni Stefchov, Galia Angelova, Preslav Nakov

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

We present the construction of an annotated corpus of PubMed abstracts reporting about positive, negative or neutral effects of treatments or substances. Our ultimate goal is to annotate one sentence (rationale) for each abstract and to use this resource as a training set for text classification of effects discussed in PubMed abstracts. Currently, the corpus consists of 750 abstracts. We describe the automatic processing that supports the corpus construction, the manual annotation activities and some features of the medical language in the abstracts selected for the annotated corpus. It turns out that recognizing the terminology and the abbreviations is key for determining the rationale sentence. The corpus will be applied to improve our classifier, which currently has accuracy of 78.80% achieved with normalization of the abstract terms based on UMLS concepts from specific semantic groups and an SVM with a linear kernel. Finally, we discuss some other possible applications of this corpus.

Original languageEnglish
Title of host publicationArtificial Intelligence
Subtitle of host publicationMethodology, Systems, and Applications - 18th International Conference, AIMSA 2018, Proceedings
EditorsJosef van Genabith, Gennady Agre, Thierry Declerck
PublisherSpringer Verlag
Pages115-125
Number of pages11
ISBN (Print)9783319993430
DOIs
Publication statusPublished - 1 Jan 2018
Event18th International Conference on Artificial Intelligence: Methodology, Systems, and Applications, AIMSA 2018 - Varna, Bulgaria
Duration: 12 Sep 201814 Sep 2018

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume11089 LNAI
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Other

Other18th International Conference on Artificial Intelligence: Methodology, Systems, and Applications, AIMSA 2018
CountryBulgaria
CityVarna
Period12/9/1814/9/18

Fingerprint

Terminology
Classifiers
Semantics
Processing
Abbreviation
Text Classification
Normalization
Annotation
Corpus
Classifier
kernel
Resources
Term

Keywords

  • Annotated rationales
  • Automatic discovery of effects
  • Corpus construction
  • PubMed abstracts
  • Terminology identification
  • Text classification

ASJC Scopus subject areas

  • Theoretical Computer Science
  • Computer Science(all)

Cite this

Stefchov, E., Angelova, G., & Nakov, P. (2018). Towards constructing a corpus for studying the effects of treatments and substances reported in pubMed abstracts. In J. van Genabith, G. Agre, & T. Declerck (Eds.), Artificial Intelligence: Methodology, Systems, and Applications - 18th International Conference, AIMSA 2018, Proceedings (pp. 115-125). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 11089 LNAI). Springer Verlag. https://doi.org/10.1007/978-3-319-99344-7_11

Towards constructing a corpus for studying the effects of treatments and substances reported in pubMed abstracts. / Stefchov, Evgeni; Angelova, Galia; Nakov, Preslav.

Artificial Intelligence: Methodology, Systems, and Applications - 18th International Conference, AIMSA 2018, Proceedings. ed. / Josef van Genabith; Gennady Agre; Thierry Declerck. Springer Verlag, 2018. p. 115-125 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 11089 LNAI).

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Stefchov, E, Angelova, G & Nakov, P 2018, Towards constructing a corpus for studying the effects of treatments and substances reported in pubMed abstracts. in J van Genabith, G Agre & T Declerck (eds), Artificial Intelligence: Methodology, Systems, and Applications - 18th International Conference, AIMSA 2018, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 11089 LNAI, Springer Verlag, pp. 115-125, 18th International Conference on Artificial Intelligence: Methodology, Systems, and Applications, AIMSA 2018, Varna, Bulgaria, 12/9/18. https://doi.org/10.1007/978-3-319-99344-7_11
Stefchov E, Angelova G, Nakov P. Towards constructing a corpus for studying the effects of treatments and substances reported in pubMed abstracts. In van Genabith J, Agre G, Declerck T, editors, Artificial Intelligence: Methodology, Systems, and Applications - 18th International Conference, AIMSA 2018, Proceedings. Springer Verlag. 2018. p. 115-125. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). https://doi.org/10.1007/978-3-319-99344-7_11
Stefchov, Evgeni ; Angelova, Galia ; Nakov, Preslav. / Towards constructing a corpus for studying the effects of treatments and substances reported in pubMed abstracts. Artificial Intelligence: Methodology, Systems, and Applications - 18th International Conference, AIMSA 2018, Proceedings. editor / Josef van Genabith ; Gennady Agre ; Thierry Declerck. Springer Verlag, 2018. pp. 115-125 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
@inproceedings{47c80f1fdbfd4863858d242808177254,
title = "Towards constructing a corpus for studying the effects of treatments and substances reported in pubMed abstracts",
abstract = "We present the construction of an annotated corpus of PubMed abstracts reporting about positive, negative or neutral effects of treatments or substances. Our ultimate goal is to annotate one sentence (rationale) for each abstract and to use this resource as a training set for text classification of effects discussed in PubMed abstracts. Currently, the corpus consists of 750 abstracts. We describe the automatic processing that supports the corpus construction, the manual annotation activities and some features of the medical language in the abstracts selected for the annotated corpus. It turns out that recognizing the terminology and the abbreviations is key for determining the rationale sentence. The corpus will be applied to improve our classifier, which currently has accuracy of 78.80{\%} achieved with normalization of the abstract terms based on UMLS concepts from specific semantic groups and an SVM with a linear kernel. Finally, we discuss some other possible applications of this corpus.",
keywords = "Annotated rationales, Automatic discovery of effects, Corpus construction, PubMed abstracts, Terminology identification, Text classification",
author = "Evgeni Stefchov and Galia Angelova and Preslav Nakov",
year = "2018",
month = "1",
day = "1",
doi = "10.1007/978-3-319-99344-7_11",
language = "English",
isbn = "9783319993430",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
publisher = "Springer Verlag",
pages = "115--125",
editor = "{van Genabith}, Josef and Gennady Agre and Thierry Declerck",
booktitle = "Artificial Intelligence",

}

TY - GEN

T1 - Towards constructing a corpus for studying the effects of treatments and substances reported in pubMed abstracts

AU - Stefchov, Evgeni

AU - Angelova, Galia

AU - Nakov, Preslav

PY - 2018/1/1

Y1 - 2018/1/1

N2 - We present the construction of an annotated corpus of PubMed abstracts reporting about positive, negative or neutral effects of treatments or substances. Our ultimate goal is to annotate one sentence (rationale) for each abstract and to use this resource as a training set for text classification of effects discussed in PubMed abstracts. Currently, the corpus consists of 750 abstracts. We describe the automatic processing that supports the corpus construction, the manual annotation activities and some features of the medical language in the abstracts selected for the annotated corpus. It turns out that recognizing the terminology and the abbreviations is key for determining the rationale sentence. The corpus will be applied to improve our classifier, which currently has accuracy of 78.80% achieved with normalization of the abstract terms based on UMLS concepts from specific semantic groups and an SVM with a linear kernel. Finally, we discuss some other possible applications of this corpus.

AB - We present the construction of an annotated corpus of PubMed abstracts reporting about positive, negative or neutral effects of treatments or substances. Our ultimate goal is to annotate one sentence (rationale) for each abstract and to use this resource as a training set for text classification of effects discussed in PubMed abstracts. Currently, the corpus consists of 750 abstracts. We describe the automatic processing that supports the corpus construction, the manual annotation activities and some features of the medical language in the abstracts selected for the annotated corpus. It turns out that recognizing the terminology and the abbreviations is key for determining the rationale sentence. The corpus will be applied to improve our classifier, which currently has accuracy of 78.80% achieved with normalization of the abstract terms based on UMLS concepts from specific semantic groups and an SVM with a linear kernel. Finally, we discuss some other possible applications of this corpus.

KW - Annotated rationales

KW - Automatic discovery of effects

KW - Corpus construction

KW - PubMed abstracts

KW - Terminology identification

KW - Text classification

UR - http://www.scopus.com/inward/record.url?scp=85053149465&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85053149465&partnerID=8YFLogxK

U2 - 10.1007/978-3-319-99344-7_11

DO - 10.1007/978-3-319-99344-7_11

M3 - Conference contribution

SN - 9783319993430

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 115

EP - 125

BT - Artificial Intelligence

A2 - van Genabith, Josef

A2 - Agre, Gennady

A2 - Declerck, Thierry

PB - Springer Verlag

ER -