Blockchain-Enabled AI-Driven Big Data Analytics for Secure and Scalable IoT Ecosystems

Abstract

The rapid evolution of Internet of Things (IoT) ecosystems produces large quantities of heterogeneous data streams, resulting in major issues related to security, scalability, privacy, and real-time data analytics. This paper suggests a novel framework of a blockchain-based and AI-oriented big data analytics system that supports secure and scalable IoT-based smart grid environments. The proposed architecture combines data acquisition using IoT, big data processing using a distributed system, intelligent predictions using machine learning techniques such as LightGBM, Random Forest, and XGBoost, and integrity preservation using a blockchain-based system, such as Hyperledger Fabric. In the proposed system, the Energy Efficiency Scores are predicted using advanced regression techniques, and LightGBM performs better with the least MAE of 0.4512, MSE of 0.325, RMSE of 0.5701, and the highest R² of 0.9988. To guarantee data immutability and trust, prediction records are cryptographically hashed using SHA-256 hashing and stored on a blockchain ledger, while sensitive payloads are kept off-chain for privacy preservation purposes. Experimental results show that accuracy is improved, residual variance is reduced, and system stability is improved. This proposed framework successfully leverages decentralized trust, intelligent analytics, and scalable processing to deliver a powerful tool for secure, transparent, and privacy-preserving IoT-based energy management systems.