Fraud Detection in Banking Data by Machine Learning Techniques

Abstract

As technology advanced and e-commerce services expanded, credit cards became one of the most popular payment methods, resulting in an increase in the volume of banking transactions. Furthermore, the significant increase in fraud requires high banking transaction costs. As a result, detecting fraudulent activities has become a fascinating topic. In this study, we consider the use of class weighttuning hyperparameters to control the weight of fraudulent and legitimate transactions. We use Bayesian optimization in particular to optimize the hyperparameters while preserving practical issues such as unbalanced data. We propose weight-tuning as a pre-process for unbalanced data, as well as CatBoost and XGBoost to improve the performance of the LightGBM method by accounting for the voting mechanism. Finally, in order to improve performance even further, we use deep learning to fine-tune the hyperparameters, particularly our proposed weight-tuning one. We perform some experiments on real-world data to test the proposed methods. To better cover unbalanced datasets, we use recall-precision metrics in addition to the standard ROC-AUC. CatBoost, LightGBM, and XGBoost are evaluated separately using a 5-fold cross-validation method. Furthermore, the majority voting ensemble learning method is used to assess the performance of the combined algorithms. LightGBM and XGBoost achieve the best level criteria of ROC-AUC = 0.95, precision 0.79, recall 0.80, F1 score 0.79, and MCC 0.79, according to the results. By using deep learning and the Bayesian optimization method to tune the hyperparameters, we also meet the ROC-AUC = 0.94, precision = 0.80, recall = 0.82, F1 score = 0.81, and MCC = 0.81. This is a significant improvement over the cutting-edge methods we compared it to.

Machine Learning Models for Fraud Detection: A Comprehensive Review and Empirical Analysis

“…- The proliferation of credit card usage in e-commerce has led to an increase in fraudulent activities, resulting in the need for effective fraud detection methods, as discussed in [5]. This study introduces weight-tuning hyperparameters, Bayesian optimization, and ensemble learning techniques to enhance fraud detection using machine learning models such as CatBoost, LightGBM, XGBoost, and deep learning.…”

Modeling the Dynamic Behaviors of Bank Account Fraudsters Using Combined Simultaneous Game Theory with Neural Networks

“…The model is put up against other classi ers like NB, SVM, and ANN and is discovered to offer effective results with high accuracy. The study [24][25][26] suggests using Bayesian optimization and class weight-tuning to identify fraudulent features in credit card transactions using gradient boosting and category boosting as an ensemble approach, the authors found Bayesian optimization to be e cient in improving computational time and performance. Additionally, deep learning [24] is used to optimize the hyperparameters and get better outcomes.…”

Presenting a Model to Detect the Fraud in Banking using Smart Enabling Tools

“…Then, if anomalous behavior or suspicious transactions are detected, the system activates an alert. To review and verify the transactions, the bank or customer can receive this alert as a text message or notification (11)(12)(13)(14)(15).…”

“…11,776 transactions is entered as input to the LightGBM algorithm, and again the supervised learning steps are performed on the training data of the filtered dataset, and then data testing is performed on all 11,776 transactions, and finally the results are according to the matrix The confusion of Figure4is obtained.…”