What question did this study set out to answer?

This research aims to develop a Quantum Federated Learning framework to enhance real-time consumer sentiment analysis while ensuring data privacy.

July 21, 2026International Journal of Academic Research in Business and Social Sciences0 citationsOpen Access

Quantum Federated Learning Framework for Privacy-Preserving Real-Time Consumer Sentiment Analysis

HYHong Zhen Yu HAHazirah Bee Binti Yusof AliCity University College of Science and Technology MCM. Kazem ChamranCity University College of Science and Technology

Key Points

This research aims to develop a Quantum Federated Learning framework to enhance real-time consumer sentiment analysis while ensuring data privacy.
Developed a hybrid model combining quantum-enhanced transformers and federated learning.
Conducted experimental analyses on datasets from IMDB, Sentiment140, Amazon, Yelp, and proprietary sources.
Evaluated performance metrics including precision, recall, and F1 score across multiple clients.
QFL outperformed traditional FL-BERT and FL-LSTM models in precision, recall, F1 score, and convergence rate.
Demonstrated significant cost and privacy savings in communication at low costs.
Maintained scalability to one thousand clients.

Abstract

HRMARS - This paper introduces a new Quantum Federated Learning (QFL) system that combines quantum-enhanced transformers and federated learning to make it possible to perform real-time consumer sentiment analysis securely, at scale, and efficiently. The hybrid model proposed overcomes three significant issues: it can be accurate when working with non-IID data, it is communication-efficient, and it can protect data privacy with high security provided by quantum cryptographic protocols. Experimental analyses on IMDB, Sentiment140, Amazon, Yelp, and our own proprietary data showed that QFL is always better in precision, recall, F1 score, and convergence rate than traditional FL-BERT and FL-LSTM models. Further, QFL was able to demonstrate substantial cost and privacy leakage savings in communication at a low cost as well as maintain scalability to one thousand clients. The study fills the gap between concepts in quantum computing and federated learning and provides a practical route to privacy-sensitive distributed intelligence on a practical scale applied to e-commerce and social media analytics.

问 AI

Bookmark

View Full Paper