Enhancing Smart Contract Access Control via Digital Identity Management and Machine Learning

In the field of blockchain technology, ensuring secure and efficient access control for smart contracts remains a critical challenge. Traditional methods are often complex and resource-intensive, potentially hindering widespread adoption. This study proposes a novel machine learning-based approach to enhance access control mechanisms. Specifically, we classify users as either benign or potentially malicious based on transaction behavior and interaction patterns. A Support Vector Machine (SVM) classifier, combined with a Genetic Algorithm (GA) for dimensionality reduction, is applied to a dataset containing ۵۰,۰۰۰ transaction records from ۱,۰۰۰ blockchain addresses. The model achieved an accuracy of ۹۴% on the test set and effectively distinguished users based on server interaction frequency and connection duration. Through visual analysis and comprehensive evaluation, we demonstrate that the proposed method improves both anomaly detection and operational efficiency. This approach has the potential to bolster trust and facilitate broader adoption of blockchain-based applications.