Title:
NEXT-GEN CLOUD SYNERGY: TRANSFORMING FINANCE, HEALTH, AND RETAIL THROUGH UNIFIED DIGITAL FRAMEWORKS

Authors:
Priyadarshini Radhakrishnan and R. Hemnath, USA

Abstract:
The most rapidly developing smart technologies in healthcare, banking, and cybersecurity need a converged, scalable, and privacy-aware intelligent system infrastructure. This paper presents a hybrid model that integrates cloudlet-based mobile cloud computing, artificial intelligence, and federated learning to address the limitations of traditional centralized models. Cloudlets offer lowlatency processing close to the edge, and AI models like XGBoost and LSTM execute prediction analytics and anomaly detection across domains. Federated learning maintains data privacy by decentralized model training, which prevents the need to centralize sensitive information. Datasets for three key domains are utilized in the research: Electronic Health Records in medicine, Credit Card Fraud Detection in finance, and CICIDS 2017 for security. Data preprocessing is performed using normalization by Min-Max and Z-score methods to offer consistent input to machine learning models. IoT and 5G technologies offer real-time data transmission and responsiveness, while Simulation-as-a-Service is used for system performance analysis in various stress conditions. Integration is maintained through standardized APIs and secure communication protocols, enabling cross-domain interoperability. Experimental deployment was realized using Python with model training and testing on the scikit-learn, TensorFlow, and PyTorch environments. The XGBoost classifier realized 98.4% accuracy in the fraud detection set, while LSTM realized 96.1% accuracy in anomaly detection in the cybersecurity set. The proposed architecture was found robust for all domains and validated its functionality in enabling smart, real-time, and secure operations in mass-scale smart systems. This strategy paves the way for future deployment of integrated smart environments in industries.