Bridging Law and Machine Learning: A Cybersecure Model for Classifying Digital Real Estate Contracts in the Metaverse

Abstract

The metaverse indicates an ever-evolving digital ecosystem where virtual real estate has now become an asset class. These properties, subject to smart contracts on the blockchain and represent as non-fungible tokens (NFTs), gives rise to new legal and cyber issues due to the decentralized and dematerialized nature of these digital assets .This paper proposes a machine learning approach to classify the digital real estate contracts into Ownership and Lease contracts. The study utilizes a dataset of one thousand digital real estate contracts collected from platforms such as Decentraland and The Sandbox. The dataset also included attributes such as plot size, plot location, transaction value, and contract duration. Preprocessing of data included encoding categorical data, standardization of numerical variables, and UTF-8 encoded text to preserve data quality. Two classification models were used: Logistic Regression and Random Forest. The model's evaluation used accuracy, precision, recall, and F1-score as evaluation criteria. The Random Forest outperformed with a perfect classification score showing that it may have been better suited to dealing with the complexity and dimensionality of the dataset. The outcomes of the study highlight the role AI could play in automating the analysis of contracts, at the same time highlighting that cybersecurity practices are important when working with data. The framework of this study seeks to support the development of a regulatory regime and add further transparency to real estate contracts in the metaverse - as a scalable tool for future digital real estate management.