Federated Predictive Learning with Privacy-Aware Model Aggregation for Distributed Analytics

Abstract

The rapid growth of IoT, edge computing, healthcare, and cyber-physical systems has increased the need for privacy-preserving distributed analytics. Traditional centralized machine learning requires sharing raw data, creating privacy, regulatory, and communication challenges. Federated Learning (FL) enables collaborative model training without exchanging sensitive data but faces limitations such as Non-IID data, communication overhead, privacy leakage, malicious updates, and inefficient aggregation. To address these issues, this study proposes a Federated Predictive Learning with Privacy-Aware Model Aggregation (FPL-PAMA) framework. The framework combines secure local training, privacy-aware weighted aggregation, client trust evaluation, and adaptive optimization to improve prediction accuracy, privacy protection, and communication efficiency. It is suitable for applications including healthcare, IoT, smart manufacturing, transportation, and financial fraud detection, providing a secure and scalable solution for next-generation distributed intelligent systems.