Detecting Smart Ponzi Schemes on Blockchain Using Machine Learning: A Comprehensive Survey

Abstract

Ponzi schemes, a more than a century-old fraud, have recently infiltrated blockchain-based cryptocurrency domain led by an explosion of such schemes in two most popular cryptocurrencies: Bitcoin and Ethereum. On these two platforms alone, the perpetrators of these frauds have fleeced gullible investors of billions of dollars annually. Smart Ponzi schemes are a hazard to these cryptocurrency ecosystems, diminishing investor confidence in these cutting-edge technologies, threatening their integrity, and hindering their growth and broader adaptation. These smart Ponzi schemes have also created a nightmare for law enforcement as tracking and taking countermeasures against fraudsters and recovering the victims’ investment is challenging. Over the years, researchers have utilized significant advances in machine learning and AI to detect and promptly caution users against investing in Ponzi schemes on Bitcoin and Ethereum. However, this research still exists in silos, and there is a lack of a detailed survey paper critically analyzing various aspects of the approaches focusing on the menace of smart Ponzi schemes. This article surveys the state-of-the-art techniques proposed in the literature to detect smart Ponzi schemes on two popular blockchain platforms: Bitcoin and Ethereum. We list, categorize, and discuss papers that contributed benchmark datasets, developed novel features concerning various aspects of smart Ponzi schemes, and proposed novel machine-learning approaches to detect them.