High Impact Factor : 4.396 icon | Submit Manuscript Online icon |

An Improved Machine Learning Method for Efficient Fraud Detection from CC Database

Author(s):

Dr. Umesh Kumar Lilhore , NRI; Vipul Patil, NRI

Keywords:

SVM, GA, Improved SVM, Credit Card, Fraud Detection, Machine Learning Method

Abstract

Credit card fraud detection methods are widely used for CC fraud detections. These techniques are based on data mining, artificial intelligence and machine learning methods. In this research work we are presenting, an improved machine learning hybrid method for efficient fraud detection from cc database. Proposed CC Fraud detection method uses the quality of improved support vector machine algorithm with k-NN and genetic algorithm. A SVM method is widely used for data classification due to its quality of results and Genetic method is used to select the best attributes set by applying best evolutionary method. Once data set is reduced by GA, SVM method is apply to classify the resulting patterns. Existing SVM method is improved by modification in Gaussian kernel parameters to improve the mapping between fraud class and normal class. Improvements are made to the hybrid with the aid of the use of a correlation degree between attributes as a health degree to update the weaker participants within the populace with newly lengthy-established chromosomes. This injects more range and will growth the overall fitness of the population. Existing k-NN method is used for best feature selection with GA. Existing SVM method and proposed ISVMGA are implemented over MATLAB Simulator and various performance measuring parameters are calculated such as accuracy, precision, sensitivity, specificity, MCC and BCR.

Other Details

Paper ID: IJSRDV6I40825
Published in: Volume : 6, Issue : 4
Publication Date: 01/07/2018
Page(s): 1006-1009

Article Preview

Download Article