A Cross-Domain Joint Probability Fusion Framework for Aspect-Based Sentiment Analysis in Special Education Contexts.

Saved in:

Bibliographic Details
Title:	A Cross-Domain Joint Probability Fusion Framework for Aspect-Based Sentiment Analysis in Special Education Contexts.
Authors:	Agrawal, Monika¹ monika.agrawal1986@gmail.com, Kumar, T. Pavan² pavankumar_ist@kluniversity.in, Rao, Moparthi Nageswara³ mnrphd@gmail.com
Source:	International Journal of Special Education. 2026 Special Issue, Vol. 41, p1-19. 19p.
Subject Terms:	*Special education, Sentiment analysis, Data fusion (Statistics), Transformer models, Language models, Graphical modeling (Statistics)
Abstract:	The particularities of the special education setting place Aspect-Based Sentiment Analysis (ABSA) under particular challenging circumstances because of the specifics of the domain, nonhomogeneous data types, and subtle emotional manifestations by students, teachers, and parents. Conventional single domain sentiment models do not account for cross domain dependencies and therefore, they do not have high interpretability and prediction accuracy. The paper solves the given problem by suggesting a Cross-Domain Joint Probability Fusion Framework (CD-JPFF) that is aimed at combining linguistic, behavioural, and contextual sentimental indicators in various modalities. The goal is to increase the accuracy of sentiment classification and maintain the level of granularity on the aspect level in delicate learning conditions. The suggested approach presents the new fusion mechanism known as Adaptive Joint Probability Sentiment Fusion (AJPSF), which generates results of transformer-based language models, behavioural feature extractors, and contextual embedding's and fuses them with the help of probabilistic graphical modelling. An architecture which makes use of hybrid architecture, based on BERT-based encoding, attention as well as Bayesian fusion is established. The algorithm calculates joint posterior probabilities in domains and optimizes them with the help of the iterative expectationmaximization. Customized test on multi-source special education data illustrates the development of considerable performance. The performance of the proposed framework is higher (in terms of accuracy, precision, recall, and F1-score) by 18.7, 16.4, 17.9, and 17.2 respectively compared to the baseline models (SVM, CNN, standalone BERT). Also, noisy and sparse data strength enhances by 21.3%. The results have shown that cross-domain probabilistic fusion is a model that is effective to model latent sentiment dependencies and enhance interpretability. The paper concludes that CD-JPFF is a scalable, reliable, context-sensitive sentiment analysis tool in the domain of special education, and it can be applied in both the personalised learning analytics and inclusive decision support systems [ABSTRACT FROM AUTHOR]
	Copyright of International Journal of Special Education is the property of International Journal of Special Education and its content may not be copied or emailed to multiple sites without the copyright holder's express written permission. Additionally, content may not be used with any artificial intelligence tools or machine learning technologies. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)
Database:	Education Research Complete

Description
Abstract:	The particularities of the special education setting place Aspect-Based Sentiment Analysis (ABSA) under particular challenging circumstances because of the specifics of the domain, nonhomogeneous data types, and subtle emotional manifestations by students, teachers, and parents. Conventional single domain sentiment models do not account for cross domain dependencies and therefore, they do not have high interpretability and prediction accuracy. The paper solves the given problem by suggesting a Cross-Domain Joint Probability Fusion Framework (CD-JPFF) that is aimed at combining linguistic, behavioural, and contextual sentimental indicators in various modalities. The goal is to increase the accuracy of sentiment classification and maintain the level of granularity on the aspect level in delicate learning conditions. The suggested approach presents the new fusion mechanism known as Adaptive Joint Probability Sentiment Fusion (AJPSF), which generates results of transformer-based language models, behavioural feature extractors, and contextual embedding's and fuses them with the help of probabilistic graphical modelling. An architecture which makes use of hybrid architecture, based on BERT-based encoding, attention as well as Bayesian fusion is established. The algorithm calculates joint posterior probabilities in domains and optimizes them with the help of the iterative expectationmaximization. Customized test on multi-source special education data illustrates the development of considerable performance. The performance of the proposed framework is higher (in terms of accuracy, precision, recall, and F1-score) by 18.7, 16.4, 17.9, and 17.2 respectively compared to the baseline models (SVM, CNN, standalone BERT). Also, noisy and sparse data strength enhances by 21.3%. The results have shown that cross-domain probabilistic fusion is a model that is effective to model latent sentiment dependencies and enhance interpretability. The paper concludes that CD-JPFF is a scalable, reliable, context-sensitive sentiment analysis tool in the domain of special education, and it can be applied in both the personalised learning analytics and inclusive decision support systems [ABSTRACT FROM AUTHOR]
ISSN:	08273383