Smart education for people with disabilities (PwDs): conceptual framework for PwDs emotions classification from student utterances (SUs) during online learning

Anwar, Aamir, Rehman, Ikram Ur ORCID: and Husamaldin, Laden (2022) Smart education for people with disabilities (PwDs): conceptual framework for PwDs emotions classification from student utterances (SUs) during online learning. In: 2022 IEEE International Smart Cities Conference (ISC2), 26-29 Sep 2022, Paphos, Cyprus.

Full text not available from this repository.


People with disabilities (PwDs) encounter numerous educational challenges, including discrimination from peers, lack of trained teachers and effective feedback mechanisms during learning. These challenges become more significant during online learning, where PwDs struggle to interact with the teachers and ask for help or provide feedback related to the content taught and the teaching strategy of the teacher. For effective feedback, emotions play a significant role in any form of interaction, and an online learning platform must be able to react to PwDs' emotional states during learning and interaction sessions. This research aims to develop a meta-emotionally intelligent model for PwDs by enhancing the feedback mechanism of online learning platforms. For this purpose, we have proposed a conceptual framework of the meta-emotional state model that classifies students' emotions using Student Utterances (SUs) collected through Google Meet sessions. This study implemented three supervised Machine Learning classification approaches, i.e., Support Vector Machine (SVM), Random Forest (RF), and AdaBoost, to classify emotions against SUs into 12 different emotion classes. Furthermore, the study experimented with 70 students with a Computer Science degree to analyse the model proficiency. The ML classifiers, SVM, provided 89%, RF 61%, and AdaBoost provided 68% accuracy while classifying SUs. For cross-validation of the classifiers' accuracy, comparative analysis through students' feedback and annotation of the sentences was conducted. The comparative analysis results show that 81.5% of the labels predicted by the ML classifiers were the same as the emotional state annotated by the students. In contrast, 11.73% of the predictions did not match the annotated results.

Item Type: Conference or Workshop Item (Paper)
ISSN: 2687-8860
ISBN: 9781665485616
Identifier: 10.1109/isc255366.2022.9922083
Page Range: pp. 1-7
Identifier: 10.1109/isc255366.2022.9922083
Keywords: Support vector machines, Radio frequency, Computer science, Analytical models, Smart cities, Annotations, Computational modeling
Subjects: Computing > Intelligent systems
Education > Teaching and learning > Technology-enhanced learning
Related URLs:
Depositing User: Ikram Rehman
Date Deposited: 07 Jun 2023 15:31
Last Modified: 21 Jun 2023 14:47

Actions (login required)

View Item View Item