Semnan University Press Journal of Modeling in Engineering 2008-4854 Articles in Press 2026 05 17 Vulnerable Smart Contracts Detection on the Ethereum Blockchain Using an Effective Risk Estimation Metric Vulnerable Smart Contracts Detection on the Ethereum Blockchain Using an Effective Risk Estimation Metric 10622 10.22075/jme.2025.36606.2793 FA Mahmood Deypir Computer Engineering 0000-0002-9417-9018 Journal Article 2025 01 17 Identifying vulnerable smart contracts has a direct impact on blockchain security because it helps users avoid using these contracts. Calculating the risk level is preferable to accurately identifying these types of contracts using classification models because these models have classification errors. In addition, the data required to achieve high accuracy may not be available. Therefore, with the help of a vulnerability risk estimation criterion for smart contracts, users can be helped in decision-making. In this research, the issue of vulnerability risk in smart contracts is introduced. In addition, an effective criterion for its estimation is devised. In this criterion, linear discriminant analysis of smart contracts and distances to their nearest neighbors is exploited to estimate the risk of an unknown smart contract. Although deep learning is not used in the proposed criterion and it requires little training data, it provides a realistic risk estimate. Experiments conducted on a real dataset of Ethereum blockchain smart contracts, including vulnerable and secure contracts, demonstrate the effectiveness of the proposed criterion. Furthermore, the performance of the proposed measure in term of detection rate, accuracy, recall and F1-score is superior to existing risk estimation metrics in other areas, such as apps and URLs. Identifying vulnerable smart contracts has a direct impact on blockchain security because it helps users avoid using these contracts. Calculating the risk level is preferable to accurately identifying these types of contracts using classification models because these models have classification errors. In addition, the data required to achieve high accuracy may not be available. Therefore, with the help of a vulnerability risk estimation criterion for smart contracts, users can be helped in decision-making. In this research, the issue of vulnerability risk in smart contracts is introduced. In addition, an effective criterion for its estimation is devised. In this criterion, linear discriminant analysis of smart contracts and distances to their nearest neighbors is exploited to estimate the risk of an unknown smart contract. Although deep learning is not used in the proposed criterion and it requires little training data, it provides a realistic risk estimate. Experiments conducted on a real dataset of Ethereum blockchain smart contracts, including vulnerable and secure contracts, demonstrate the effectiveness of the proposed criterion. Furthermore, the performance of the proposed measure in term of detection rate, accuracy, recall and F1-score is superior to existing risk estimation metrics in other areas, such as apps and URLs.