Comparison of anomaly credit card scam discovery using naive bayes algorithm with decision tree algorithm with improved accuracy

• Published in: AIP conference proceedings Vol. 2821; no. 1
• Main Authors: Kumar, K. Yashwanth, Vani, B.
• Format: Journal Article Conference Proceeding
• Language: English
• Published: Melville American Institute of Physics 21.11.2023
• Subjects: Algorithms; Charge exchange; Decision trees; Mathematical analysis
• ISSN: 0094-243X; 1551-7616
• DOI: 10.1063/5.0177044

This study aims to identify the credit card fraud using the Naive Bayes calculation and contrast its performance with Decision tree algorithm. The example size for extortion location for exact oddity discovery with further developed quality was test N=20 (Group 1=10 cycles and Group 2=10 emphasis) is taken for the two calculations. In the dataset 492, extortion exchanges are available out of general exchanges. Gullible Bayes calculation is contrasted and different AI calculations utilizing measurements like exactness, particularity and f-score. Results: The discovery correctness’s of Naive Bayes calculations and Decision tree are 97.6%, and 93.7% separately. It shows the factual importance P=0.001 (P<0.05) 2-followed by a free example T-test Conclusion: Results exhibit that the proposed gullible Bayes calculation gives altogether preferable execution over the Decision tree calculations in imaginative extortion discovery for fake exchanges in charge cards.