The E-Wallets service market is more and more interested by many business companies particular to the banks. In Vietnam, this service can be seen as the objective for a potential card-less environment of payments. Although the service deployments are encouraged by many procedures from Government, there is still a limited number of E-Wallets‟ customers in the payment market. Some reasons are an inconvenient transaction infrastructure, or Vietnamese behaviors, etc. But the main problem leading to the reluctance of using E-wallet is security issue on online transactions such as personal information losses, or fake bank account or fraudulent transaction, etc. The advantage of using machine learning in transaction process can be seen as a good solution for securing customers‟ online payments. Machine learning classifiers such as Support Vector Machine, Radial Basis Function, and Decision Tree can help transaction systems to recognize fraudulent payments. This can be performed by classifiers‟ prediction from historical data. In this paper, a potential E-Wallet loan model using machine learning algorithms in Peer to Peer (P2P) business systems for Vietnamese Banks is shown. The dataset is taken from Vietnamese financial companies for experiments with alternative machine learning algorithms. The MASI dataset (data is re-organized with a reduction of its imbalance via algorithm) is also shown with better results.
Volume 12 | 07-Special Issue
Pages: 1636-1641
DOI: 10.5373/JARDCS/V12SP7/20202268