Privacy Preserving Topic-wise Sentiment Analysis of the Iran Israel USA Conflict Using Federated Transformer Models
arXiv:2603.13655v1 Announce Type: new Abstract: The recent escalation of the Iran Israel USA conflict in 2026 has triggered widespread global discussions across social media platforms. As people increasingly use these platforms for expressing opinions, analyzing public sentiment from these discussions can provide valuable insights into global public perception. This study aims to analyze global public sentiment regarding the Iran Israel USA conflict by mining user-generated comments from YouTube news channels. The work contributes to public opinion analysis by introducing a privacy preserving framework that combines topic wise sentiment analysis with modern deep learning techniques and Federated Learning. To achieve this, approximately 19,000 YouTube comments were collected from major international news channels and preprocessed to remove noise and normalize text. Sentiment labels were initially generated using the VADER sentiment analyzer and later validated through manual inspection
arXiv:2603.13655v1 Announce Type: new Abstract: The recent escalation of the Iran Israel USA conflict in 2026 has triggered widespread global discussions across social media platforms. As people increasingly use these platforms for expressing opinions, analyzing public sentiment from these discussions can provide valuable insights into global public perception. This study aims to analyze global public sentiment regarding the Iran Israel USA conflict by mining user-generated comments from YouTube news channels. The work contributes to public opinion analysis by introducing a privacy preserving framework that combines topic wise sentiment analysis with modern deep learning techniques and Federated Learning. To achieve this, approximately 19,000 YouTube comments were collected from major international news channels and preprocessed to remove noise and normalize text. Sentiment labels were initially generated using the VADER sentiment analyzer and later validated through manual inspection to improve reliability. Latent Dirichlet Allocation (LDA) was applied to identify key discussion topics related to the conflict. Several transformer-based models, including BERT, RoBERTa, XLNet, DistilBERT, ModernBERT, and ELECTRA, were fine tuned for sentiment classification. The best-performing model was further integrated into a federated learning environment to enable distributed training by preserving user data privacy. Additionally, Explainable Artificial Intelligence (XAI) techniques using SHAP were applied to interpret model predictions and identify influential words affecting sentiment classification. Experimental results demonstrate that transformer models perform effectively, and among them, ELECTRA achieved the best performance with 91.32% accuracy. The federated learning also maintained strong performance while preserving privacy, achieving 89.59% accuracy in a two client configuration.
Executive Summary
This article proposes a novel approach to analyzing global public sentiment regarding the Iran Israel USA conflict by leveraging Federated Learning and modern deep learning techniques. The authors introduce a privacy-preserving framework that combines topic-wise sentiment analysis with transformer-based models. Experimental results demonstrate the effectiveness of this approach, with ELECTRA achieving 91.32% accuracy in sentiment classification. The federated learning environment maintains strong performance while preserving user data privacy. This study contributes to public opinion analysis and has significant implications for practical and policy applications.
Key Points
- ▸ The authors propose a novel approach to analyzing global public sentiment using Federated Learning and modern deep learning techniques.
- ▸ A privacy-preserving framework is introduced that combines topic-wise sentiment analysis with transformer-based models.
- ▸ ELECTRA achieves 91.32% accuracy in sentiment classification, outperforming other transformer-based models.
Merits
Strength in Federated Learning
The authors successfully integrate Federated Learning with transformer-based models to preserve user data privacy while maintaining strong performance.
Effective Topic-wise Sentiment Analysis
The proposed approach effectively captures topic-wise sentiment analysis by leveraging Latent Dirichlet Allocation (LDA) and transformer-based models.
Explainable Artificial Intelligence (XAI)
The authors apply SHAP-based XAI techniques to interpret model predictions and identify influential words affecting sentiment classification.
Demerits
Limited Dataset
The study relies on a relatively small dataset of approximately 19,000 YouTube comments, which may not be representative of the broader global public opinion.
Limited Generalizability
The results may not be generalizable to other conflict scenarios or social media platforms due to the specific focus on the Iran Israel USA conflict and YouTube news channels.
Expert Commentary
The article presents a novel approach to analyzing global public sentiment using Federated Learning and modern deep learning techniques. While the results are promising, the study's limitations, such as the limited dataset and generalizability, need to be addressed. The proposed approach has significant implications for practical and policy applications, particularly in conflict resolution and public diplomacy. Furthermore, the use of XAI techniques provides valuable insights into the models' predictions and identifies influential words affecting sentiment classification. However, the study's dependence on YouTube news channels and the specific focus on the Iran Israel USA conflict may limit its generalizability. Nevertheless, the authors' successful integration of Federated Learning with transformer-based models is a significant contribution to the field of public opinion analysis.
Recommendations
- ✓ Future studies should focus on expanding the dataset and generalizability of the proposed approach to other conflict scenarios and social media platforms.
- ✓ The authors should explore the application of the proposed approach in real-world conflict scenarios to further validate its effectiveness.