Citation

Sungheetha, Akey and Saranya, S. Lakshmi (2026) AI-Powered Sustainable Healthcare Analytics Platform for Cancer Prediction with Privacy-Preserving Data Visualization and Green Computing Integration. In: 7th International Conference on Mobile Computing and Sustainable Informatics, ICMCSI 2026, 7 January 2026 - 9 January 2026, Goathgaun.

Text IEEE Xplore Full-Text PDF_.pdf - Published Version Restricted to Repository staff only Download (366kB)

Abstract

Cancer remains a leading global health challenge with approximately 19.3 million new cases diagnosed annually, demanding innovative computational approaches for early detection and sustainable healthcare delivery. Traditional cancer prediction systems face critical limitations including high computational energy consumption, privacy vulnerabilities in patient data handling, and inadequate real-time analytics capabilities for clinical decision-making. This research addresses these challenges by developing an AI-Powered Sustainable Healthcare Analytics Platform that integrates privacy-preserving mechanisms with green computing principles for cancer prediction and visualization. The proposed methodology combines federated learning architectures with homomorphic encryption techniques, achieving 94.7% prediction accuracy while reducing energy consumption by 67.3 % compared to centralized cloud-based approaches. The platform implements differential privacy protocols with epsilon values of 0.85, ensuring patient data confidentiality while maintaining statistical utility for machine learning models. Novel contributions include the development of an Adaptive Energy-Aware Neural Architecture (AENA) that dynamically adjusts computational complexity based on available renewable energy sources, a Privacy-Preserving Visualization Framework (PPVF) enabling secure multi-institutional data sharing, and a Green Computing Optimization Module (GCOM) reducing carbon footprint by 58.4 % through intelligent workload scheduling. Experimental validation using publicly available cancer datasets demonstrates superior performance with F1-scores of 0.932, specificity of 91.8 %, and processing throughput of 1847 records per second while consuming only 34.2 watts average power. The system achieves real-time prediction latency of 127 milliseconds, making it suitable for clinical deployment in resource-constrained healthcare environments seeking sustainable digital transformation.

Item Type:	Conference or Workshop Item (Paper)
Uncontrolled Keywords:	Machine learning
Subjects:	Q Science > Q Science (General) > Q300-390 Cybernetics
Divisions:	Faculty of Information Science and Technology (FIST)
Depositing User:	Ms Rosnani Abd Wahab
Date Deposited:	04 May 2026 01:54
Last Modified:	04 May 2026 01:54
URII:	http://shdl.mmu.edu.my/id/eprint/15818

Downloads

Edit (login required)