SECURE LAND RECORD MANAGEMENT THROUGH ARTIFICIAL INTELLIGENCE AND DECENTRALIZED BLOCKCHAIN ARCHITECTURE

Abstract

The rapid digitalization of land registration processes has heightened apprehensions over data security, transparency, and fraud detection. Centralized databases used by traditional land registration systems are susceptible to data tampering, illegal access, cyberattacks, and fraudulent activities. To rectify these shortcomings, the developers of this initiative present GREENLAND, an intelligent and secure land registration system using Blockchain, IPFS, ML, and XAI. The proposed system guarantees immutability, transparency, and decentralized governance by securely archiving user data and land transaction records on the Ethereum blockchain via the use of smart contracts. To safeguard vast land-related records and minimize storage requirements, IPFS archives them with their associated hash values documented on the blockchain. A variety of machine learning models are created and assessed on a dataset of real transactions to identify fraud. The models include Logistic Regression, Support Vector Machine (SVM), Random Forest, XGBoost, and LightGBM. LightGBM outperforms its competitors with an accuracy of 99.35%. Utilizing SHAP and LIME to clarify fraud predictions bolsters confidence and improves interpretability for human understanding. A scalable, safe, transparent, and contemporary system for digital land registration and governance is built via the combination of Blockchain, IPFS, machine learning, and explainable artificial intelligence.