期刊名称：International Journal of Innovative Research in Computer and Communication Engineering
印刷版ISSN：2320-9798
电子版ISSN：2320-9801
出版年度：2017
卷号：5
期号：9
页码：15521
DOI：10.15680/IJIRCCE.2017.0509022
出版社：S&S Publications
摘要：The number of Credit Card transactions has been astonishingly rising and due to this breathtakingproblem, bunch of users are trapped in Credit Card Fraud. So, up gradation of existing Credit Card Fraud Detectiontechniques has become a major priority for the banks to minimize their daily losses. In this paper, the researcher hasused the concept of Support Vector Machine methodology along with the usage of Specialized Kernels to detect thefrauds. The Kernel Trick is a class of algorithms used for pattern analysis, whose best known member is the supportvector machine (SVM). The objective of pattern analysis is to find and study general types of relations (for exampleclusters, correlations, classifications) in datasets. This methodology has been proved to be one of the finest andmagnificent approaches in fraud identification because the accuracy of the system has been tested against Linear, RBFand Polynomial Kernel. As a result, polynomial kernel shows the maximum accuracy in the output.
关键词：Credit Card Discovery; Credit Card Fraud Detection; Kernel Trick; Linear Kernel; RBF Kernel;Polynomial Kernel; Support Vector Machine
