Modular Federated Cross-Domain Recommendation (MFCDR) System with a Projected Attention Network

Manisha Shrirang Otari; Dipika Patil; Mithun Basawaraj Patil

doi:10.47852/bonviewJCCE52026111

Authors

Manisha Shrirang Otari Department of Computer Science and Engineering, Sanjay Ghodawat University, India https://orcid.org/0000-0002-9759-6706
Dipika Patil Department of Computer Science and Engineering, Sanjay Ghodawat University, India
Mithun Basawaraj Patil Department of Computer Science and Engineering, Nagesh Karajagi Orchid College of Engineering & Technology, India https://orcid.org/0000-0001-7741-3645

DOI:

https://doi.org/10.47852/bonviewJCCE52026111

Keywords:

federated learning, attention network, privacy, cross-domain recommendation, data security

Abstract

In today's digital age, data is considered a new currency. It drives many aspects related to research, business strategies, and decisions across various industries, providing recommendations. As data becomes more valuable, the privacy of user data becomes crucial. This article introduces an innovative, novel privacy-preserving federated learning approach, with the advancement of attention networks for cross-domain recommendation. A decentralized approach to federated learning has replaced traditional machine learning to enhance user privacy and data security. The research employs a projected attention network (PRADO) within a real federated environment to improve local device training. The proposed framework is examined through a use case of book recommendation based on emotions extracted from social media reviews, utilizing GoEmotions and customized book dataset. The results showed assurance of the system with better performance in terms of precision, recall, and the F1-score over benchmarking models such as Embedding and Mapping framework for Cross-Domain Recommendation (EMCDR), Cross-Domain based on Latent Feature Mapping (CDFLM), Aspect-based Neural Recommender (ANR), and Meta-learning-based model for Federated Personalized Cross-Domain Recommendation (MFPCDR), with the suggested system reaching a precision of 0.96 and an F1-score of 0.89. The findings indicate that the real implementation of a federated learning environment is modular, scalable, efficient, and preserves privacy.

Received: 8 May 2025 | Revised: 31 July 2025 | Accepted: 13 September 2025

Conflicts of Interest

The authors declare that they have no conflicts of interest to this work.

Data Availability Statement

Data are available from the corresponding author upon reasonable request.

Author Contribution Statement

Manisha Shrirang Otari: Conceptualization, Methodology, Software, Validation, Formal analysis, Investigation, Data curation, Writing – original draft, Writing – review & editing, Visualization, Project administration. Dipika Patil: Validation, Formal analysis, Supervision. Mithun Basawaraj Pati: Software, Validation, Formal analysis, Investigation, Data curation, Writing – original draft, Writing – review & editing, Visualization, Supervision, Project administration.