Evaluation of Attention-Based LSTM and Bi-LSTM Networks For Abstract Text Classification in Systematic Literature Review Automation

Regina Ofori-Boateng; Magaly Aceves-Martins; Chrisina Jayne; Nirmalie Wiratunga; Carlos Francisco Moreno-Garcia

doi:10.1016/j.procs.2023.08.149

Evaluation of Attention-Based LSTM and Bi-LSTM Networks For Abstract Text Classification in Systematic Literature Review Automation

Regina Ofori-Boateng^*, Magaly Aceves-Martins, Chrisina Jayne, Nirmalie Wiratunga, Carlos Francisco Moreno-Garcia

^*Corresponding author for this work

The Rowett Institute of Nutrition and Health

Research output: Contribution to journal › Conference article › peer-review

Abstract

Systematic Review (SR) presents the highest form of evidence in research for decision and policy-making. Nonetheless, the structured steps involved in carrying out SRs make it demanding for reviewers. Many studies have projected the abstract screening stage in the SR process to be the most burdensome for reviewers, thus automating this stage with artificial intelligence (AI). However, majority of these studies focus on using traditional machine learning classifiers for the abstract classification. Thus, there remain a gap to explore the potential of deep learning techniques for this task. This study seeks to bridge the gap by exploring how LSTM and Bi-LSTM models together with GloVe for vectorisation can accelerate this stage. As a further aim to increase precision while sustaining a recall >= 95% due to precision-recall trade-off, attention mechanics is added to these classifiers. The final experimental results obtained showed that Bi-LSTM with attention has the capacity to expedite citation screening.

Original language	English
Pages (from-to)	114-126
Number of pages	13
Journal	Procedia Computer Science
Volume	222
Early online date	31 Aug 2023
DOIs	https://doi.org/10.1016/j.procs.2023.08.149
Publication status	Published - 2023
Event	International Neural Network Society Workshop on Deep Learning Innovations and Applications, INNS DLIA 2023 - Gold Coast, Australia Duration: 18 Jun 2023 → 23 Jun 2023

Bibliographical note

Funding Information:
The authors would like to thank members of the COMO project 6 for supporting this research.

Keywords

abstract screening
artificial intelligence
Bi-LSTM
deep learning
LSTM
machine learning
Systematic literature review

Access to Document

10.1016/j.procs.2023.08.149Licence: CC BY-NC-ND

Ofori-Boateng_etal_PCS_Evaluation_of_Attention-Based_VOR
This is an open access article under the CC BY-NC-ND license (https://creativecommons.org/licenses/by-nc-nd/4.0)
Final published version, 605 KBLicence: CC BY-NC-ND

Cite this

@article{9d37f24f44074a639757691695a50f52,

title = "Evaluation of Attention-Based LSTM and Bi-LSTM Networks For Abstract Text Classification in Systematic Literature Review Automation",

abstract = "Systematic Review (SR) presents the highest form of evidence in research for decision and policy-making. Nonetheless, the structured steps involved in carrying out SRs make it demanding for reviewers. Many studies have projected the abstract screening stage in the SR process to be the most burdensome for reviewers, thus automating this stage with artificial intelligence (AI). However, majority of these studies focus on using traditional machine learning classifiers for the abstract classification. Thus, there remain a gap to explore the potential of deep learning techniques for this task. This study seeks to bridge the gap by exploring how LSTM and Bi-LSTM models together with GloVe for vectorisation can accelerate this stage. As a further aim to increase precision while sustaining a recall >= 95% due to precision-recall trade-off, attention mechanics is added to these classifiers. The final experimental results obtained showed that Bi-LSTM with attention has the capacity to expedite citation screening.",

keywords = "abstract screening, artificial intelligence, Bi-LSTM, deep learning, LSTM, machine learning, Systematic literature review",

author = "Regina Ofori-Boateng and Magaly Aceves-Martins and Chrisina Jayne and Nirmalie Wiratunga and Moreno-Garcia, {Carlos Francisco}",

note = "Funding Information: The authors would like to thank members of the COMO project 6 for supporting this research. ; International Neural Network Society Workshop on Deep Learning Innovations and Applications, INNS DLIA 2023 ; Conference date: 18-06-2023 Through 23-06-2023",

year = "2023",

doi = "10.1016/j.procs.2023.08.149",

language = "English",

volume = "222",

pages = "114--126",

journal = "Procedia Computer Science",

issn = "1877-0509",

publisher = "Elsevier BV",

}

TY - JOUR

T1 - Evaluation of Attention-Based LSTM and Bi-LSTM Networks For Abstract Text Classification in Systematic Literature Review Automation

AU - Ofori-Boateng, Regina

AU - Aceves-Martins, Magaly

AU - Jayne, Chrisina

AU - Wiratunga, Nirmalie

AU - Moreno-Garcia, Carlos Francisco

N1 - Funding Information: The authors would like to thank members of the COMO project 6 for supporting this research.

PY - 2023

Y1 - 2023

N2 - Systematic Review (SR) presents the highest form of evidence in research for decision and policy-making. Nonetheless, the structured steps involved in carrying out SRs make it demanding for reviewers. Many studies have projected the abstract screening stage in the SR process to be the most burdensome for reviewers, thus automating this stage with artificial intelligence (AI). However, majority of these studies focus on using traditional machine learning classifiers for the abstract classification. Thus, there remain a gap to explore the potential of deep learning techniques for this task. This study seeks to bridge the gap by exploring how LSTM and Bi-LSTM models together with GloVe for vectorisation can accelerate this stage. As a further aim to increase precision while sustaining a recall >= 95% due to precision-recall trade-off, attention mechanics is added to these classifiers. The final experimental results obtained showed that Bi-LSTM with attention has the capacity to expedite citation screening.

AB - Systematic Review (SR) presents the highest form of evidence in research for decision and policy-making. Nonetheless, the structured steps involved in carrying out SRs make it demanding for reviewers. Many studies have projected the abstract screening stage in the SR process to be the most burdensome for reviewers, thus automating this stage with artificial intelligence (AI). However, majority of these studies focus on using traditional machine learning classifiers for the abstract classification. Thus, there remain a gap to explore the potential of deep learning techniques for this task. This study seeks to bridge the gap by exploring how LSTM and Bi-LSTM models together with GloVe for vectorisation can accelerate this stage. As a further aim to increase precision while sustaining a recall >= 95% due to precision-recall trade-off, attention mechanics is added to these classifiers. The final experimental results obtained showed that Bi-LSTM with attention has the capacity to expedite citation screening.

KW - abstract screening

KW - artificial intelligence

KW - Bi-LSTM

KW - deep learning

KW - LSTM

KW - machine learning

KW - Systematic literature review

UR - http://www.scopus.com/inward/record.url?scp=85175845426&partnerID=8YFLogxK

U2 - 10.1016/j.procs.2023.08.149

DO - 10.1016/j.procs.2023.08.149

M3 - Conference article

AN - SCOPUS:85175845426

SN - 1877-0509

VL - 222

SP - 114

EP - 126

JO - Procedia Computer Science

JF - Procedia Computer Science

T2 - International Neural Network Society Workshop on Deep Learning Innovations and Applications, INNS DLIA 2023

Y2 - 18 June 2023 through 23 June 2023

ER -

Evaluation of Attention-Based LSTM and Bi-LSTM Networks For Abstract Text Classification in Systematic Literature Review Automation

Abstract

Bibliographical note

Keywords

Access to Document

Other files and links

Fingerprint

Cite this