A Robust Framework for fraud Detection in Banking using ML and NN

Banking fraud is a problem that is becoming more and more serious, along with considerable monetary losses, damage to the bank's brand, loss of client and customer confidence. Fraud identification and prevention are major challenges for many financial organizations, retail firms, and e-commerce...

Full description

Saved in:
Bibliographic Details
Published in:Proceedings of the National Academy of Sciences, India, Section A, physical sciences India, Section A, physical sciences, 2024-04, Vol.94 (2), p.201-212
Main Authors: Vashistha, Astha, Tiwari, Anoop Kumar, Singh, Priyanshi, Yadav, Paritosh Kumar, Pandey, Sudhakar
Format: Article
Language:English
Subjects:
Applied and Technical Physics
Artificial neural networks
Atomic
Banking
Datasets
Decision trees
Feasibility studies
Fraud
Fraud prevention
Machine learning
Molecular
Multilayer perceptrons
Neural networks
Optical and Plasma Physics
Physics
Physics and Astronomy
Quantum Physics
Scientific Research Paper
Support vector machines
Citations: Items that this one cites
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Banking fraud is a problem that is becoming more and more serious, along with considerable monetary losses, damage to the bank's brand, loss of client and customer confidence. Fraud identification and prevention are major challenges for many financial organizations, retail firms, and e-commerce companies. Fraud detection is used to both identify and stop fraudsters from obtaining goods or bugs illegally. In the same vein, this research will conduct a feasibility study to determine the best fraud detection strategy. We provide a list of the tried-and-true methods for spotting fraud. To avoid fraud detection, many techniques like Deep Neural Network, Support Vector Machine, Multilayer Perceptron, K-Nearest Neighbors, Random Forest, XG Boost, LGBM, and Decision Tree were used. The dataset was built from 20,000 entries on Kaggle, each having 114 attributes. Before using machine learning and neural network approaches, the dataset is balanced using the Synthetic Minority Over-Sampling Method. Following the analysis of the dataset using a number of methods, it was determined that Random Forest, Decision Tree, XG Boost, and LGBM all had 100% accuracy. This demonstrates that the model outperformed other models by balancing the dataset.
ISSN:0369-8203
2250-1762
DOI:10.1007/s40010-024-00871-1