Machine Learning-Based Risk Classification of Blockchain Cryptocurrency Transactions Using Ensemble Algorithms
Journal of Current Research in Blockchain · 02.11.2026 · 3 (4) · 263-278
DOI: 10.47738/jcrb.v3i4.84 ↗Kurz gefasst
Fachartikel in der Zeitschrift Journal of Current Research in Blockchain, online erschienen am 02.11.2026. Verfasst von R. Karthick Manoj.
Abstract (Originalsprache)
Blockchain cryptocurrency transactions are increasingly exposed to financial risks, fraudulent activities, and suspicious transaction behaviors due to the decentralized nature of blockchain networks. This study proposes a machine learning-based framework for blockchain cryptocurrency transaction risk classification using ensemble learning algorithms. The dataset consists of 9,000 blockchain transactions with multiple behavioral and transaction-related attributes, including transaction amount, transaction fee, wallet age, transaction velocity, darknet association score, and unusual transaction patterns. To address the class imbalance problem, the SMOTEENN resampling technique was applied, transforming the original class distribution of 6,086 normal transactions and 1,114 risky transactions into a more balanced distribution of 5,866 risky transactions and 2,588 normal transactions. Several ensemble learning algorithms, including Random Forest, Gradient Boosting, XGBoost, and LightGBM, were evaluated using confusion matrix analysis and ROC-AUC metrics. Experimental results demonstrate that XGBoost achieved the highest risky transaction detection capability with 169 True Positive predictions and 110 False Negatives, while LightGBM generated the highest True Negative value of 1,422 and the lowest False Positive value of 99. Random Forest achieved the highest ROC-AUC score of 0.553531, followed by LightGBM with 0.542258. Feature importance analysis revealed that output_count and input_count were the most influential attributes with importance scores of 1289 and 1224, respectively. Although the obtained ROC-AUC values indicate that blockchain transaction risk classification remains a challenging problem, the findings demonstrate that ensemble learning algorithms can support blockchain risk analytics and cryptocurrency fraud detection systems.
Zitieren
APA
Manoj, R. K. (2026). Machine Learning-Based Risk Classification of Blockchain Cryptocurrency Transactions Using Ensemble Algorithms. Journal of Current Research in Blockchain, 3(4), 263-278. https://doi.org/10.47738/jcrb.v3i4.84
Vancouver
Manoj RK. Machine Learning-Based Risk Classification of Blockchain Cryptocurrency Transactions Using Ensemble Algorithms. Journal of Current Research in Blockchain. 2026;3(4):263-278. doi:10.47738/jcrb.v3i4.84
BibTeX
@article{Manoj2026,
author = {Manoj, R. Karthick},
title = {Machine Learning-Based Risk Classification of Blockchain Cryptocurrency Transactions Using Ensemble Algorithms},
journal = {Journal of Current Research in Blockchain},
year = {2026},
volume = {3},
number = {4},
pages = {263--278},
doi = {10.47738/jcrb.v3i4.84}
}
EUVIMED