EUVIMEDEUVIMED entdecken
Crossref

An Explainable Machine Learning Framework for Proxy Smart Contract Identification in Ethereum

Hasan Aladwan

Journal of Current Research in Blockchain · 02.11.2026 · 3 (4) · 248-262

DOI: 10.47738/jcrb.v3i4.83 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Journal of Current Research in Blockchain, online erschienen am 02.11.2026. Verfasst von Hasan Aladwan.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Proxy smart contracts play a critical role in enabling upgradeability and maintainability within blockchain-based applications. However, identifying proxy contracts from large-scale smart contract repositories remains a challenging task due to the diversity of implementation patterns and the increasing complexity of Solidity source code. This study proposes an explainable machine learning framework for Ethereum proxy smart contract identification using source code analysis and SHAP-based interpretability. A dataset consisting of 576 validated Ethereum smart contracts, including 23 proxy contracts and 553 non-proxy contracts, was constructed through a comprehensive data cleaning and preprocessing process. Solidity source code was transformed into numerical representations using the Term Frequency–Inverse Document Frequency (TF-IDF) technique, while Synthetic Minority Over-sampling Technique (SMOTE) was applied to address class imbalance. Four machine learning algorithms, namely Logistic Regression, Random Forest, Support Vector Machine (SVM), and XGBoost, were evaluated and compared. Experimental results demonstrated that Random Forest achieved the best overall performance with an accuracy of 99.14%, a precision of 83.33%, a recall of 100%, an F1-score of 90.91%, and an AUC of 99.64%. Furthermore, SHAP analysis revealed that ERC1967-related patterns, implementation references, and beacon-based upgradeability mechanisms were the most influential indicators for proxy contract detection. The findings indicate that the proposed framework not only achieves excellent predictive performance but also provides transparent and interpretable explanations for classification decisions. This study contributes to blockchain security research by offering an effective and explainable approach for automated proxy smart contract identification.

Zitieren

APA

Aladwan, H. (2026). An Explainable Machine Learning Framework for Proxy Smart Contract Identification in Ethereum. Journal of Current Research in Blockchain, 3(4), 248-262. https://doi.org/10.47738/jcrb.v3i4.83

Vancouver

Aladwan H. An Explainable Machine Learning Framework for Proxy Smart Contract Identification in Ethereum. Journal of Current Research in Blockchain. 2026;3(4):248-262. doi:10.47738/jcrb.v3i4.83

BibTeX
@article{Aladwan2026,
  author = {Aladwan, Hasan},
  title = {An Explainable Machine Learning Framework for Proxy Smart Contract Identification in Ethereum},
  journal = {Journal of Current Research in Blockchain},
  year = {2026},
  volume = {3},
  number = {4},
  pages = {248--262},
  doi = {10.47738/jcrb.v3i4.83}
}
Kostenlos starten →