Analysis of Credit Card Fraud Data Using Various Machine Learning Methods
C Chethana, Piyush Kumar Pareek
Abstract
The credit card details of customers are very important. Due to online transactions, fraud is also increasing by those who have unauthorized access. Various machine learning algorithms are used to perform an analysis on the credit card fraud information. The analysis of the transaction details can be done during the transaction, or the details can be fetched from stored databases. In this chapter, different tree methods are used to perform the analysis and evaluate the performance. An accuracy of about 99.9% without Principal Component Analysis (PCA) and 99.8% of accuracy with PCA is obtained. The analysis is performed using various algorithms like the K-nearest neighbor method (KNN), the support vector machine (SVM) method, the Gaussian naive Bayesian algorithm, and the logistic regression (LR) algorithm. The analysis was done using the option of principal component analysis and without principal component analysis with good results of about 99.9% accuracy obtained for the SVM, KNN, and LR algorithms.