Enhancing Predictive Digital Health Analytics Using Blockchain-Based Graph Intelligence
Pertanika Journal of Artificial Intelligence · 30.10.2026 · 1 (1)
DOI: 10.47836/pjai.1.1.04 ↗Kurz gefasst
Fachartikel in der Zeitschrift Pertanika Journal of Artificial Intelligence, online erschienen am 30.10.2026. Verfasst von Gunasekar Thangarasu und 1 weiteren Autor:innen.
Themen: Digitalisierung in der Pflege
Abstract (Originalsprache)
Digital health systems have faced persistent challenges in predictive reliability, data integrity, and secure interoperability across distributed stakeholders. Conventional machine learning models have relied on centralized data repositories, which have limited trust, auditability, and longitudinal consistency. Graph-theoretical approaches have emerged as promising tools for modeling complex health interactions, while blockchain architecture has ensured immutability and decentralized trust. However, its combined application for health prediction has remained underexplored. Conventional digital health prediction frameworks have lacked a unified mechanism that captures the relational dependencies among patients, symptoms, and clinical outcomes while simultaneously guaranteeing secure data provenance. Most of the graph-based models have ignored trust management, and most of the blockchain-based systems have underutilized structural intelligence, which has resulted in fragmented and inefficient predictive pipelines. This study has proposed a novel graph-theoretic algorithm named Adaptive Health Influence Graph Traversal (AHIGT), which has modeled patient-centric health events as weighted temporal graphs. The algorithm has computed predictive risk scores through an influence propagation that has adapted to the dynamic edge relevance. A permissioned blockchain layer has been combined to store graph updates, model parameters, and prediction hashes, which has ensured transparency and tamper resistance. Smart contracts have governed data access and model update validation, which has preserved privacy while enabling collaborative learning. Experimental evaluation on benchmark digital health datasets has shown that the proposed AHIGT framework has achieved improved predictive performance with a maximum accuracy of 0.85, precision of 0.84, recall of 0.83, and an F1-score of 0.84 compared with conventional graph-based methods. Computation time per patient remains efficient at 1.05 seconds, which shows scalability for real-time deployment. The addition of adaptive graph traversal with a blockchain has allowed data integrity, auditability, and collaborative updates, confirming the robustness and reliability of the predictive framework under dynamic and multi-source health data conditions.
Zitieren
APA
Thangarasu, G., & Alla, K. R. (2026). Enhancing Predictive Digital Health Analytics Using Blockchain-Based Graph Intelligence. Pertanika Journal of Artificial Intelligence, 1(1). https://doi.org/10.47836/pjai.1.1.04
Vancouver
Thangarasu G, Alla KR. Enhancing Predictive Digital Health Analytics Using Blockchain-Based Graph Intelligence. Pertanika Journal of Artificial Intelligence. 2026;1(1). doi:10.47836/pjai.1.1.04
BibTeX
@article{Thangarasu2026,
author = {Thangarasu, Gunasekar and Alla, Kesava Rao},
title = {Enhancing Predictive Digital Health Analytics Using Blockchain-Based Graph Intelligence},
journal = {Pertanika Journal of Artificial Intelligence},
year = {2026},
volume = {1},
number = {1},
doi = {10.47836/pjai.1.1.04}
}
EUVIMED