Identifying likely reputable ethereum blockchain based projects

Abstract

The identification of likely‐reputable projects based on the Ethereum Blockchain is a crucial task in the ever‐evolving landscape of blockchain technology. As the number of Ethereum‐based projects increases, differentiating between trustworthy ventures and potentially fraudulent ones becomes increasingly challenging. While previous research has focused on detecting illicit behavior, it has not addressed the specific aspect of investigating whether a project’s potential to be a reputable one could be determined in an automated fashion. Consequently, this research aims to investigate whether machine learning algorithms and data analysis techniques can be used to assist investors, developers, regulators, and the Ethereum community in evaluating project reputations. The research endeavors to achieve classification of a dataset consisting of 2,179 accounts identified by the Ethereum community for their association with illicit activities, alongside 3,977 accounts that are considered to have a high likelihood of being reputable, obtained from CoinGecko. Through evaluation process utilizing 10‐fold cross‐validation, the LightGBM algorithm achieves an impressive average accuracy of 0.9984 (±0.0003) and an average AUC of 0.9991 (±0.0003). Notably, the most influential features in the final model are “Time difference between received transactions (minutes),” and “Average time difference between ERC20 received transactions.” Based on the obtained results, it was shown that the proposed approach effectively identifies likely reputable projects on the Ethereum Blockchain. The outcomes of this research will contribute to establishing a more informed and secure investment environment within the Ethereum ecosystem. Stakeholders will benefit from making well‐informed decisions, mitigating risks associated with unreliable projects, and fostering the growth of reputable ventures. Additionally, this research sets the groundwork for future advancements in project evaluation methodologies, further enhancing the development and maturity of the Ethereum Blockchain ecosystem.