ENHANCING FRAUD DETECTION IN FINANCIAL TRANSACTIONS: A COMPARATIVE ANALYSIS OF MACHINE LEARNING ALGORITHMS

Abstract

In the modern financial landscape, fraudulent personal information in financial transactions threatens data integrity and security. There is potential for machine learning to identify and mitigate fraud threats. This comparative study examines machine learning techniques for detecting and preventing the use of fraudulent personal information in financial transactions. Using a large dataset of financial transaction records, logistic regression, k-nearest neighbors, decision trees, gradient boosting, and ensemble methods such as Adaboost, XGboost, LightGBM, and CatBoost are evaluated. Presented is a comprehensive analysis and comparison of each algorithm's ability to detect and eradicate false personal information. The findings highlight the significance of using the optimal algorithm to optimize the detection and prevention of financial transaction fraud. The study also sheds light on the interpretability and scalability of the algorithms, allowing for the incorporation of robust and flexible machine learning approaches in financial security and data protection. In conclusion, this study offers crucial insights regarding the selection and application of machine learning algorithms to combat the use of fraudulent personal information in financial transactions. The findings highlight the need for sophisticated machine learning solutions to combat financial crime and strengthen data integrity and security standards.