TY  - JOUR
Y1  - 2025///
EP  -  668
SP  - 653 
TI  - Dynamic Fraud Detection in E-banking: A Case Study on Ethereum Networks Using Machine Learning Enhanced by Blockchain
JF  - Lecture Notes in Electrical Engineering
PB  - Springer Science and Business Media Deutschland GmbH
N1  - Cited by: 0
A1  - Maheshwari, Vikash Chander
A1  - Kumar, Ganesh
A1  - Osman, Nurul Aida
A1  - Singh, Ranjit
A1  - Imam, Abdullahi Abubakar
A1  - Hoong, Angela Lee Siew
A1  - Hajjami, Salma El
ID  - scholars20561
KW  - Banking; Crime; Data privacy; Learning systems; Machine learning; Network security; Random forests; Block-chain; Blockchain technology; Case-studies; Cybe security fintech; Cyber security; E-banking; Financial fraud; Fraud detection; Machine-learning; Random forests; Blockchain
AV  - none
N2  - The occurrence of financial fraud is increasing despite advancements in technology. The unavailability of reliable trustworthy financial transaction data due to a lack of inter-organizational collaboration and privacy concerns poses challenge for data-driven technologies such as ML, which require reliable data for effective performance in real-world applications. This research presents the use of machine learning and blockchain technology to develop a robust fraud detection model for e-banking. This study focuses on addressing the issues of fraudulent activities and unusual events within the Ethereum network. The suggested approach employs blockchain to protect the confidentiality of data, which is widely recognized as a highly secure technology in the finance sector. Machine learning algorithms are employed in conjunction with blockchain technology to identify fraudulent transactions on the Ethereum network. There are two algorithms employed for transaction classification which are XGBoost and Random Forest (RF), while GridSearchCV is utilized for hyperparameter tuning, approach involves utilizing XGBoost, random forest, and logistics regression algorithms for classifying transactions and predicting patterns of transactions. This is accomplished by training the dataset with both fraudulent and integrated transaction patterns, and then applying machine learning techniques to predict new incoming transactions. To assess the accuracy of the models, we will determine their precision and AUC values. © Institute of Technology PETRONAS Sdn Bhd (Universiti Teknologi PETRONAS) 2025.
UR  - https://www.scopus.com/pages/publications/105012923940?origin=resultslist
VL  - 1417 L
SN  - 18761100
ER  -