A hybrid medical text classification framework: integrating attentive rule construction and neural network

The main objective of this work is to improve the quality and transparency of the medical text classification solutions. Conventional text classification methods provide users with only a restricted mechanism (based on frequency) for selecting features. In this paper, a three-stage hybrid method com...

Full description

Bibliographic Details
Main Authors: Li, Xiang, Cui, Menglin, Li, Jingpeng, Bai, Ruibin, Lu, Zheng, Aickelin, Uwe
Format: Article
Language:English
Published: 2021
Subjects:
Online Access:https://eprints.nottingham.ac.uk/65138/
_version_ 1848800191889539072
author Li, Xiang
Cui, Menglin
Li, Jingpeng
Bai, Ruibin
Lu, Zheng
Aickelin, Uwe
author_facet Li, Xiang
Cui, Menglin
Li, Jingpeng
Bai, Ruibin
Lu, Zheng
Aickelin, Uwe
author_sort Li, Xiang
building Nottingham Research Data Repository
collection Online Access
description The main objective of this work is to improve the quality and transparency of the medical text classification solutions. Conventional text classification methods provide users with only a restricted mechanism (based on frequency) for selecting features. In this paper, a three-stage hybrid method combining the threshold-gated attentive bi-directional Long Short-Term Memory (ABLSTM) and the regular expression based classifier is proposed for medical text classification tasks. The bi-directional Long Short-Term Memory (LSTM) architecture with an attention layer allows the network to weigh words according to their perceived importance and focus on crucial parts of a sentence. Feature words (or keywords) extracted by ABLSTM model are utilized to guide the regular expression rule construction. Our proposed approach leverages the advantages of both the interpretability of rule-based algorithms and the computational power of deep learning approaches for a production-ready scenario. Experimental results on real-world medical online query data clearly validate the superiority of our system in selecting domain-specific and topic-related features. Results show that the proposed approach achieves an accuracy of 0.89 and an F1-score of 0.92 respectively. Furthermore, our experimentation also illustrates the versatility of regular expressions as a user-level tool for focusing on desired patterns and providing interpretable solutions for human modification.
first_indexed 2025-11-14T20:47:39Z
format Article
id nottingham-65138
institution University of Nottingham Malaysia Campus
institution_category Local University
language English
last_indexed 2025-11-14T20:47:39Z
publishDate 2021
recordtype eprints
repository_type Digital Repository
spelling nottingham-651382021-05-12T02:54:56Z https://eprints.nottingham.ac.uk/65138/ A hybrid medical text classification framework: integrating attentive rule construction and neural network Li, Xiang Cui, Menglin Li, Jingpeng Bai, Ruibin Lu, Zheng Aickelin, Uwe The main objective of this work is to improve the quality and transparency of the medical text classification solutions. Conventional text classification methods provide users with only a restricted mechanism (based on frequency) for selecting features. In this paper, a three-stage hybrid method combining the threshold-gated attentive bi-directional Long Short-Term Memory (ABLSTM) and the regular expression based classifier is proposed for medical text classification tasks. The bi-directional Long Short-Term Memory (LSTM) architecture with an attention layer allows the network to weigh words according to their perceived importance and focus on crucial parts of a sentence. Feature words (or keywords) extracted by ABLSTM model are utilized to guide the regular expression rule construction. Our proposed approach leverages the advantages of both the interpretability of rule-based algorithms and the computational power of deep learning approaches for a production-ready scenario. Experimental results on real-world medical online query data clearly validate the superiority of our system in selecting domain-specific and topic-related features. Results show that the proposed approach achieves an accuracy of 0.89 and an F1-score of 0.92 respectively. Furthermore, our experimentation also illustrates the versatility of regular expressions as a user-level tool for focusing on desired patterns and providing interpretable solutions for human modification. 2021-03-10 Article PeerReviewed application/pdf en cc_by https://eprints.nottingham.ac.uk/65138/1/A%20hybrid%20medical%20text%20classification%20framework%EF%BC%9A%20Integrating%20attentive%20rule%20construction%20and%20neural%20network.pdf Li, Xiang, Cui, Menglin, Li, Jingpeng, Bai, Ruibin, Lu, Zheng and Aickelin, Uwe (2021) A hybrid medical text classification framework: integrating attentive rule construction and neural network. Neurocomputing, 443 . pp. 345-355. ISSN 09252312 hybrid system; deep learning; attention mechanism; text classification http://dx.doi.org/10.1016/j.neucom.2021.02.069 doi:10.1016/j.neucom.2021.02.069 doi:10.1016/j.neucom.2021.02.069
spellingShingle hybrid system; deep learning; attention mechanism; text classification
Li, Xiang
Cui, Menglin
Li, Jingpeng
Bai, Ruibin
Lu, Zheng
Aickelin, Uwe
A hybrid medical text classification framework: integrating attentive rule construction and neural network
title A hybrid medical text classification framework: integrating attentive rule construction and neural network
title_full A hybrid medical text classification framework: integrating attentive rule construction and neural network
title_fullStr A hybrid medical text classification framework: integrating attentive rule construction and neural network
title_full_unstemmed A hybrid medical text classification framework: integrating attentive rule construction and neural network
title_short A hybrid medical text classification framework: integrating attentive rule construction and neural network
title_sort hybrid medical text classification framework: integrating attentive rule construction and neural network
topic hybrid system; deep learning; attention mechanism; text classification
url https://eprints.nottingham.ac.uk/65138/
https://eprints.nottingham.ac.uk/65138/
https://eprints.nottingham.ac.uk/65138/