Credit Card Fraudulent Transactions Prediction Using Novel Sequential Transactions by Comparing Light Gradient Booster Algorithm Over Isolation Forest Algorithm

The Aim of the work is to predict the accuracy of credit card fraudulent detection using Sequential transactions by comparing Light Gradient Booster over Isolation Forest. Light Gradient Booster Algorithm per sample size of 10 and Isolation Forest (IF) with sample size of 10 was repeated for identifying the accuracy percentage of credit card fraudulent transactions. The sigmoid function used in the Light Gradient Booster Algorithm maps the values between 0 and 1. Light Gradient Booster Algorithm has better accuracy (91.6%) when compared to Isolation Forest (81.8%). There is a statistical significant difference between Light Gradient Booster Algorithm and Isolation Forest with p=0.0001 (p<0.05) based on 2-tailed analysis. Light Gradient Booster Algorithm shows a better accuracy percentage of credit card fraudulent transactions than Isolation Forest.