Blockchain is a distributed ledger that has provided transparency, immutability, and security for transactions and decentralized applications. Ethereum, in particular, is a public blockchain platform that utilizes smart contracts to execute transactions and agreements amongst parties on the network. Although Ethereum has revolutionized blockchain technology through smart contracts, the platform has also created a haven for adversaries to facilitate fraudulent transactions. Typical schemes amongst Ethereum include but are not limited to phishing scams and Ponzi schemes. This paper categorizes the most notable fraud schemes on the platform and applies machine learning (ML) to detect illicit transactions on the Ethereum blockchain. In addition, this paper also develops a new fraud triangle adapted to Ethereum to understand the motivations for committing fraud on the network. The goal is to provide a comprehensive overview of fraud in Ethereum and identify critical patterns and fraud schemes through data mining techniques and real-world scenario analysis.