No Detail Found!

ADOCA: A novel technique to defraud credit card using an optimzed catboost algorithm

Author(s)
KANEEZ ZAINAB Department of Computer Science and Engineering, Amity University, Lucknow Campus, Uttar Pradesh, India
NAMRATA DHANDA Department of Computer Science and Engineering, Amity University, Lucknow Campus, Uttar Pradesh, India
QAMAR ABBAS Department of Computer Science and Engineering, Ambalika Institute of Technology & Management, Uttar Pradesh, India
Abstract
Cashless economy has increased the demand of digital affairs. Online transactions using Credit cards is one of the most often used medium of digital transactions. Spike in recent years is seen in fraudulent transactions across the digital platform. The researchers have suggested many techniques in the past for detection of fraudulent transactions. But due to the key challenges like the changing profiles of both fraudulent and nonfraudulent transaction and data being unbalanced hinders technologies like data mining and major algorithm of machine learning (such as KNN, SVM, Random Forest and Decision Tree) and models of deep learning. Therefore, a novel proposal has been suggested for detecting the credit card fraud transaction using an optimized CatBoost Algorithm for determining that whether the transaction is legitimate or fraudulent by optimizing the Bayesian-based hyper parameter to tune the parameter of the CatBoost Algorithm. Hence, we suggest this novel approach as ADOCA (Anomaly Detection using an Optimized CatBoost Algorithm). Based on that, we compare our approach from the different binary classification algorithms that includes Logistic Regression, KNN, SVC, Decision Tree and Random Forest.
Publication Details
Page(s) 1196-1205
DOI DOI not available
Published Journal: Journal of Theoretical and Applied Information Technology, Volume: 100, Issue: 4, Year: 2022

Request the full text from PASTIC

Keywords
Machine learning, Optimized CatBoost, Credit card fraud, CatBoost, Binary Classification
0

Citations

0

Downloads

110

Views