Predictability of Metaverse Coins Using an Advanced Machine Learning Approach
摘要
Metaverse coins are a type of crypto asset used for making transactions on the Metaverse blockchain. Most of these cryptocurrencies have come into existence in the last 2–3 years and garnered investors’ attention. This chapter examines the predictability of the top four Metaverse coins based on their market capitalization—Decentraland, The Sandbox, Stacks, and Theta Network. A residual-driven two-stage random forest model is proposed for this purpose. It is an extension of the traditional random forest model. The residual of the first stage is used as an explanatory feature for the second stage of forecasting. The feature importance study confirms the capability of residues to influence the future movements of the considered Metaverse coins. The predictive modeling exercise indicates that the future movements of all four coins can be estimated with high precision based on the values of four performance measures for the test, training, and full datasets. This study's findings will help investors select appropriate investment channels for diversification and risk mitigation.