WiP: Generative Adversarial Network for Oversampling Data in Credit Card Fraud Detection

Online credit card transactions have witnessed a steady year-to-year growth, making fraud detection an increasingly critical challenge. Credit card fraud detection is particularly difficult due to the extreme class imbalance between fraudulent and non-fraudulent transactions. Imbalanced data, where one class heavily dominates the other, must be addressed before effective fraud detection can be achieved. Traditional oversampling techniques such as the Synthetic Minority Over-sampling Technique (SMOTE) and its variants often generate synthetic samples without considering the importance of ranking performance in the minority class distribution, leading to suboptimal classification performance. To address this, an adaptive Weight and rank-based Minority (WRM-SMOTE) oversampling method is proposed that prioritizes the performance and its relative ranking. To the best of our knowledge, WRM-SMOTE is among the first method to combine performance-based weighting and ranking across multiple SMOTE variants for adaptive oversampling. This method transforms imbalanced datasets into balanced ones through adaptive synthetic sample generation. The process begins by computing a weight based on two componentsabsolute performance and relative ranking. This is done to provide a distinction between two variants with extremely similar performance when one outperforms the other. Our experimental results on real-world credit card fraud datasets and eleven other benchmark datasets demonstrate that WRM-SMOTE significantly improves fraud detection performance, achieving higher precision and recall while maintaining a balanced false-positive rate compared to traditional oversampling methods. The proposed method effectively enhances classification accuracy by generating synthetic samples in high-impact regions thereby improving the robustness of fraud detection models.

Machine learning approaches are widely used to analyze and detect the increasingly serious problem of credit card fraud. However, typical credit card datasets present imbalanced classification situations because of severely skewed class distributions. Although researchers have proposed some strategies to deal with these imbalances, disadvantages remain. We propose an oversampling method based on variational automatic coding (VAE), combined with classic deep learning techniques, to solve this problem. The VAE method is used to generate a large amount of diverse cases from minority groups in an imbalanced dataset, which are then used to train the classification network. The proposed method is tested on an open credit card fraud dataset, which contains transactions conducted by European cardholders over two days in September 2013. Experimental results show that the VAE method performs better than synthetic minority oversampling techniques and traditional deep neural network methods. In addition, it outperforms recent oversampling methods based on generative adversarial network (GAN) models. After submitting the extended dataset to the baseline for training, the test of the VAE model performs well on indicators including precision, F-measure, accuracy and specificity. These experimental results suggest that the VAE-based oversampling method can be effectively applied to imbalanced classification problems.

Financial fraud is a serious threat that is expanding effects on the financial sector. The use of credit cards is growing as digitization and internet transactions advance daily. The most common issues in today's culture are credit card scams. This kind of fraud typically happens when someone uses someone else's credit card details. Credit card fraud detection uses transaction data attributes to identify credit card fraud, which can save significant financial losses and affluence the burden on the police. The detection of credit card fraud has three difficulties: uneven data, an abundance of unseen variables, and the selection of an appropriate threshold to improve the models' reliability. This study employs a modified Logistic Regression (LR) model to detect credit card fraud in order to get over the preceding difficulties. The dataset sampling strategy, variable choice, and detection methods employed all have a significant impact on the effectiveness of fraud detection in credit card transactions. The effectiveness of naive bayes, k-nearest neighbour, and logistic regression on highly skewed credit card fraud data is examined in this research. The accuracy of the logistic regression technique will be closer to 0.98%; with this accuracy, frauds may be easily detected. The fact that LR receives the highest classifier score illustrates how well LR predicts credit card theft.

The increasing number of online financial transactions, particularly those involving credit cards, underscores the urgent need for robust security systems to mitigate financial losses due to fraud. This paper presents a novel machine learning-based approach to credit card fraud detection that addresses the growing demand for enhanced security. Unlike traditional single-model approaches, the proposed ensemble model uniquely combines random forest, logistic regression, and AdaBoost techniques to accurately distinguish between legitimate and fraudulent transactions. The novelty of this work lies in its innovative soft voting mechanism, which aggregates the strengths of these diverse algorithms to achieve superior classification accuracy and minimize false positives. By leveraging the complementary strengths of each model, the ensemble approach provides a robust and adaptive framework capable of detecting emerging fraud patterns in real-time. The results demonstrate that the proposed ensemble model outperforms individual models, achieving an accuracy of 99.96%, a precision of 99.53%, and a recall of 100%, with an F1 score of 0.99 and an Area Under the Curve (AUC) of 1.0. These findings highlight the model's ability to significantly reduce false positives and negatives, making it a highly reliable solution for fraud detection. Furthermore, the model's performance was validated on the PaySim dataset, a large-scale synthetic dataset, where it achieved a 99.97% accuracy, demonstrating its generalizability and robustness in handling complex, real-world scenarios. This research contributes to the field by introducing a highly extensible and adaptable fraud detection framework that improves current solutions and provides a foundation for future advancements in ensemble learning for fraud detection. The proposed model's ability to integrate multiple classifiers while maintaining interpretability and computational efficiency sets it apart from previous studies, offering a promising direction for enhancing the security of online financial transactions.

The increase of Credit Card (CC) fraud in recent years requires the development of fraud detection systems that are both efficient and robust. This paper explored the utilization of machine learning models, with a particular emphasis on ensemble methods, to advance the detection of CC fraud. We present an ensemble model that incorporates different classifiers to address the dataset imbalance issue that is present in most CC datasets. We employed synthetic over-sampling and under-sampling techniques in certain machine learning algorithms to tackle the same issue. To address the issue of under sampling, we implemented ensemble random sampling, which was subsequently followed by Two-Stage Thresholding (TST) to eliminate outliers. The well-known Synthetic Minority Over-sampling Technique (SMOTE) was used for over-sampling. Data was utilized to train the model and subsequently generate predictions by utilizing testing data following the pre-processing of the dataset. The proposed ensemble model outperformed conventional methods in addressing the challenges associated with CC fraud detection, as demonstrated by a comparative analysis. Our findings suggest that ensemble approaches are effective at fighting fraud, paving the way for more adaptive and resilient fraud detection systems.

The existence of fraud in credit card transactions causes many financial losses leading to customers’ loss of trust. Fraud detection methods based on machine learning techniques prevent such losses. Despite the literature on fraud detection, there is a lack of algorithms that detect fraud with acceptable performance in the credit card fraud detection field. Therefore, this study proposed a comprehensive ensemble-based method using an efficient weighted voting strategy for credit card fraud detection that can address the previous algorithms’ weaknesses. First, since the dataset is imbalanced, the proposed method balanced the dataset by stratifying it into three different proportions of normal and fraudulent transactions (1 to 1, 1 to 4 and 1 to 9 ratios). The features in each dataset are ranked by four feature-ranking methods, and the Random Forest classifier is applied to each of them for selecting the essential features. Afterward, using the seven base classifiers and the chosen features, 12 ensembles have been developed. Besides, a weighted voting strategy is proposed, and the fraudulent transactions are detected through voting based on the base classifiers’ and ensembles’ weights, which are calculated by their performance. The computational results indicated that the suggested Eclf10 is the best ensemble and its Logistic Regression classifier also has the best performance among other base classifiers. The Eclf10 leads to 99.97% accuracy, 87.78% precision, 97.70% recall, 92.21% F1-score and 95.634% F2-score, which has a superiority over the previous ensemble-based methods (e.g. majority voting ensemble, stacking classifier, Adaboost, Gradient Boosting).

Cyber-security experts often require the help of an automated process that filters and classifies network attacks. To apply specific preventive measures for securing networks, the classification of the attack type is the key. Many Machine Learning (ML) models have been proposed as a base for Network Intrusion Detection (NID) systems. However, their performance varies based on multiple factors. For instance, an ML model fitted on a highly imbalanced dataset can be biased toward over-represented attack types. On the other hand, paying attention only to the ML model’s performance in the minority classes can negatively affect its performance in the majority classes. This paper proposes an NID system that addresses the issue of imbalanced datasets and uses Convolutional Neural Networks (CNN) to classify the different attack types. We compare the performance of our proposed system to other systems that use: Random Over-Sampling (ROS), Synthetic Minority Oversampling TEchnique (SMOTE), Adaptive Synthetic Sampling (ADASYN), and Generative Adversarial Networks (GAN). Using the NSL-KDD and the BoT-IoT datasets for benchmarking, we show that our proposed system performs well in the minority classes: recall scores of 70.50% and 72.08% on the User to Root (U2R) and Remote to Local (R2L) attack classes of the NSL-KDD dataset, respectively, while maintaining an overall False Alarm Rate (FAR) of 6.50% and a recall of 90.46% on the binary classification task. Our proposed system scores a weighted average F1-Score of 99.45% on the multi-class classification task using the BoT-IoT dataset.

Credit card fraud resulted in the loss of $3 billion to North American financial institutions in 2017. The rise of digital payments systems such as Apple Pay, Android Pay, and Venmo has meant that loss due to fraudulent activity is expected to increase. Deep Learning presents a promising solution to the problem of credit card fraud detection by enabling institutions to make optimal use of their historic customer data as well as real-time transaction details that are recorded at the time of the transaction. In 2017, a study found that a Deep Learning approach provided comparable results to prevailing fraud detection methods such as Gradient Boosted Trees and Logistic Regression. However, Deep Learning encompasses a number of topologies. Additionally, the various parameters used to construct the model (e.g. the number of neurons in the hidden layer of a neural network) also influence its results. In this paper, we evaluate a subsection of Deep Learning topologies — from the general artificial neural network to topologies with built-in time and memory components such as Long Short-term memory — and different parameters with regard to their efficacy in fraud detection on a dataset of nearly 80 million credit card transactions that have been pre-labeled as fraudulent and legitimate. We utilize a high performance, distributed cloud computing environment to navigate past common fraud detection problems such as class imbalance and scalability. Our analysis provides a comprehensive guide to sensitivity analysis of model parameters with regard to performance in fraud detection. We also present a framework for parameter tuning of Deep Learning topologies for credit card fraud detection to enable financial institutions to reduce losses by preventing fraudulent activity.

Credit card fraud detection is highly important to maintain financial security. However, it is challenging to train suitable models due to the class imbalance in credit card transaction data. To address this issue, this work proposes a novel deep learning framework, gated attention-based generative adversarial networks (GA-GAN) for credit card fraud detection in class-imbalanced data. GA-GAN integrates GAN and the gated attention mechanism to generate high-quality synthetic data that realistically simulates fraudulent behaviors. Experimental results on two public credit card datasets demonstrate that GA-GAN outperforms state-of-the-art methods on credit card fraud detection tasks in class-imbalanced data, indicating the advantage of GA-GAN. The code is publicly available at https://github.com/Gejiangmeng/gagan/tree/main.

Optimised autoencoder-based ensemble deep learning approaches for cyber-physical event classification utilizing synchrophasor PMU data

This paper uses a classifier named random forest to detect credit card fraud. Credit card fraud is one of the main issues in the economic industry. To construct a credit card fraud detection system, a certain amount of samples is required. In this paper, a dataset containing 284,807 credit card transactions is used. This dataset has gone through the PCA transformation and includes 492 frauds out of 284,807 transactions. Based on the huge amount of data and imbalanced samples, this paper compresses the dataset and uses the synthetic minority over-sampling technique (SMOTE) to address the problem of imbalanced samples. Also, in this paper, we use random forest as a classification model while constructing the fraud detection system/method.

With the development of the financial industry, the number of credit cards has greatly increased. However, credit card fraud is still a major concern for financial security. Credit card fraud detection is a typical imbalanced classification problem, in which fraudulent cardholders are far less than non-fraudulent cardholders. The training on imbalanced samples will cause that the classifier ignores the minor fraudulent samples. To solve this problem, a generative adversarial network (GAN) based classification method is proposed for credit card fraud detection. GAN consists of a generative model and a discriminative model, and the two models are trained in a competitive way to get the Nash equilibrium. Specifically, the generative model tries to fit the real distribution of the non-fraudulent samples. The discriminative model determines the probability of a sample belongs to the distribution of non-fraudulent samples. To improve the discrimination performance, the fraudulent samples are also used in the training of discriminative model, which is different from the traditional GAN training scheme. The experimental results show that the recall reaches 82.7%.

<p>In recent years, escalating fraudulent activities have led to significant financial losses across industries, intensifying the critical challenge of fraud detection. This study introduces a novel hybrid model that combines artificial neural networks (ANN) with support vector machines (SVM) to construct a robust additive model for fraud detection. Emphasizing the Synthetic Minority Over-sampling Technique (SMOTE), our investigation addresses the imbalanced nature of fraud versus non-fraud transactions. The clear novelty of our research lies in the seamless integration of these two powerful tools, offering a comprehensive and effective solution to the challenges posed by credit card fraud detection. Furthermore, our study stands out by emphasizing the collaborative synergy between ANN and SVM, particularly through the integration of multiple kernels, which improves the adaptability and accuracy of the proposed hybrid model. We conducted a thorough examination of 284,807 anonymized transactions, placing special emphasis on comparing the hybrid approach's performance and showcasing its superiority over traditional methodologies in the realm of fraud detection.</p>

Credit card fraud is a major problem in today’s financial world. It induces severe damage to financial institutions and individuals. There has been an exponential increase in the losses due to fraud in recent years. Hence, effectively detecting fraudulent behavior is of vital importance for either financial institutions or individuals. Since credit fraud events account for a small proportion of all transaction events in real life, the datasets about credit fraud are usually imbalanced. Some common classifiers, such as decision tree and naïve Bayes, are unable to detect fraud. Furthermore, in some cases, traditional strategies for dealing with an imbalanced problem, such as the synthetic minority oversampling technique (SMOTE), are not effective for the fraud detection datasets. To accurately detect fraud behavior, this study uses anomaly detection on imbalanced data, as well as Isolation Forest (IForest) with kernel principal component analysis. A one-class support vector machine (OCSVM) with AdaBoost is used as two models to detect outliers which significantly improves detection accuracy and efficiency. The model achieved 96% accuracy, 100% precision, 96% recall, and 98% F 1 score, respectively. The proposed model is expected to become a helpful tool for finding credit card fraud detection, and the analysis presented in this study will provides useful insights into credit card fraud detection mechanisms.

Detecting credit card fraud is a major challenge because fraudulent transactions represent only a small fraction of financial data. Traditional methods like SMOTE (Synthetic Minority Oversampling Technique) help balance datasets but can also introduce noise and make models over- fit, reducing their effectiveness. To tackle the issues that come with oversampling, we present the Enhanced Fraud Classifier with Tomek Links (EFC-Tomek) framework. This approach builds on the existing EFN-SMOTE but takes a different approach, using Tomek Links undersampling instead of SMOTE oversampling to balance the dataset. Our main goal is to improve data quality and enhance the model’s ability to detect fraud more accurately and effectively. To test EFC- Tomek, we used two real-world datasets: European cardholders and Loan Prediction. We evaluated its performance using a number of classifiers, such as Random Forest, eXtreme Gradient Boosting, Logistic Regression, Gradient Boosting, Artificial Neural Networks, and Support Vector Classifier. The results showed that EFC-Tomek improved fraud detection, with ANN achieving the highest accuracy on both datasets.