Enhancing aspect-based sentiment analysis with dependency-attention GCN and mutual assistance mechanism

Result code in IS VaVaI

<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F00216208%3A11320%2F25%3AYLV8BYFR" target="_blank" >RIV/00216208:11320/25:YLV8BYFR - isvavai.cz</a>

Result on the web

<a href="https://www.scopus.com/inward/record.uri?eid=2-s2.0-85169341824&doi=10.1007%2fs10844-023-00811-2&partnerID=40&md5=da9160014c49c7606a9df31667672baf" target="_blank" >https://www.scopus.com/inward/record.uri?eid=2-s2.0-85169341824&doi=10.1007%2fs10844-023-00811-2&partnerID=40&md5=da9160014c49c7606a9df31667672baf</a>

DOI - Digital Object Identifier

<a href="http://dx.doi.org/10.1007/s10844-023-00811-2" target="_blank" >10.1007/s10844-023-00811-2</a>

Result language

angličtina

Original language name

Enhancing aspect-based sentiment analysis with dependency-attention GCN and mutual assistance mechanism

Original language description

Aspect-based sentiment analysis (ABSA) has been extensively studied in recent years. It involves several subtasks for extracting one or more sentiment elements, including the aspect category, aspect term, opinion term, and sentiment polarity. In this paper, we propose two novel approaches to addressing different ABSA subtasks. Firstly, we introduce a Dependency-Attention GCN-based Aspect Opinion Extractor (DAG-AOE) for the Aspect-Opinion Pair Extraction (AOPE) task. DAG-AOE employs an improved graph convolutional network to extract syntactic structure information from text sequences, thereby effectively identifying aspect-opinion pairs in sentences. Secondly, we propose a Mutual Assistance Mechanism-based Category Sentiment Classifier (MAM-CSC) that utilizes the results of DAG-AOE to address the Aspect Sentiment Quad Prediction (ASQP) task. MAM-CSC leverages the semantic relationships between words in a sentence and addresses the two independent classification tasks through a mutual assistance approach. We conduct extensive experiments on benchmark datasets, and the experimental results demonstrate that our models have demonstrated a significant improvement over all baseline methods. Specifically, our models achieved the highest improvement of 1.4% F1 score over the baseline on the AOPE task and 1.5% F1 score on the ASQP task. © The Author(s), under exclusive licence to Springer Science+Business Media, LLC, part of Springer Nature 2023.

Czech name

—

Czech description

—

Type

J_SC - Article in a specialist periodical, which is included in the SCOPUS database

CEP classification

—

OECD FORD branch

10201 - Computer sciences, information science, bioinformathics (hardware development to be 2.2, social aspect to be 5.8)

Project

—

Continuities

—

Publication year

2024

Confidentiality

S - Úplné a pravdivé údaje o projektu nepodléhají ochraně podle zvláštních právních předpisů

Name of the periodical

Journal of Intelligent Information Systems

ISSN

0925-9902

e-ISSN

—

Volume of the periodical

62

Issue of the periodical within the volume

1

Country of publishing house

US - UNITED STATES

Number of pages

27

Pages from-to

163-189

UT code for WoS article

—

EID of the result in the Scopus database

2-s2.0-85169341824