NORMA eResearch @NCI Library

Detecting and Mitigating Opinion Polarization in Online Review Systems: A Multi-Modal Framework Using Graph Neural Networks and Temporal Analysis

Eddula, Aman (2025) Detecting and Mitigating Opinion Polarization in Online Review Systems: A Multi-Modal Framework Using Graph Neural Networks and Temporal Analysis. Masters thesis, Dublin, National College of Ireland.

[thumbnail of Master of Science]
Preview
PDF (Master of Science)
Download (815kB) | Preview
[thumbnail of Configuration Manual]
Preview
PDF (Configuration Manual)
Download (357kB) | Preview

Abstract

Online review platforms such as Amazon and Yelp influence millions of consumer decisions, yet they are increasingly affected by opinion polarization and echo chamber effects. This study introduces the Sentiment Echo Detection System (SEDS), a hybrid framework designed to detect and quantify polarization in user-generated reviews. Unlike traditional sentiment analysis approaches that treat reviews as independent textual units, SEDS models the review space as a semantic similarity graph and integrates contextual embeddings from lightweight transformers (MiniLM) with Graph Neural Networks (GCNs). Reviews are embedded, connected via cosine similarity, and passed through a GCN to classify sentiment while capturing relational patterns in the data. We evaluate our approach on 50,000 reviews from each platform and demonstrate competitive performance compared to baseline classifiers, with the added benefit of modeling community-driven sentiment structure. This work presents a scalable and interpretable alternative to traditional NLP methods, enabling platforms to better understand and potentially mitigate the formation of polarized sentiment clusters.

Item Type: Thesis (Masters)
Supervisors:
Name
Email
Kumar, Teerath
UNSPECIFIED
Subjects: P Language and Literature > P Philology. Linguistics > Computational linguistics. Natural language processing
Q Science > Q Science (General) > Self-organizing systems. Conscious automata > Machine learning
Z Bibliography. Library Science. Information Resources > ZA Information resources > ZA4150 Computer Network Resources > The Internet > World Wide Web > Websites
T Technology > TK Electrical engineering. Electronics. Nuclear engineering > Telecommunications > The Internet > World Wide Web > Websites
Divisions: School of Computing > Master of Science in Data Analytics
Depositing User: Ciara O'Brien
Date Deposited: 25 Aug 2026 13:33
Last Modified: 25 Aug 2026 13:33
URI: https://norma.ncirl.ie/id/eprint/9627

Actions (login required)

View Item View Item