Hybrid Undersampling and Oversampling for Handling Imbalanced Credit Card Data

Abstract

Recent developments in the use of credit cards for a range of daily life activities have increased credit card fraud and caused huge financial losses for individuals and financial institutions. Most credit card frauds are conducted online through illegal payment authorizations by data breaches, phishing, or scams. Many solutions have been suggested for this issue, but they all face the major challenge of building an effective detection model using highly imbalanced class data. Most sampling techniques used for class imbalance have limitations, such as overlapping and overfitting, which cause inaccurate learning and are slowed down by noisy features. Herein, a hybrid Tomek links BIRCH clustering borderline SMOTE (BCB-SMOTE) sampling method is proposed to balance a highly skewed credit card transaction dataset. First, Tomek links were used to undersample majority instances and remove noise, and then BIRCH clustering was applied to cluster the data and oversample minority instances using B-SMOTE. The credit card frauddetection model was run using a random forest (RF) classifier. The proposed method achieved a higher F1score (85.20%) than the baseline sampling techniques tested for comparison. Because of the enormous number of credit card transactions, there was still a small false-positive rate. The proposed method improves the detection performance owing to the well-organized balancing of the dataset.

Exploring Resampling Techniques in Credit Card Default Prediction

“…The Tomek links undersampling (TLUS) method is used to balance imbalanced datasets by removing observations located between two different classes. The basic idea behind Tomek links is to identify pairs of observations, one from the majority class and one from the minority class, that are closest to each other but belong to different classes (Alamri & Ykhlef, 2024;Devi et al, 2017). These pairs are called Tomek links (Devi et al, 2017, p. 3).…”

“…While TomekLinks undersampling resulted in higher overall accuracy compared to RUS and CNN undersampling techniques, the latter two tend to have more consistent and moderate precision and recall scores. These ndings indicate that RUS and CNN may provide a more balanced trade-off between precision, recall, and accuracy, whereas TomekLinks may prioritize accuracy at the expense of precision and recall(Alamri & Ykhlef, 2024;Devi et al, 2017). Furthermore, the fact that precision and recall scores are consistent across different algorithms suggests that RUS and CNN may be able to perform better in a variety of situations compared to TomekLinks.In terms of the AUC, TomekLinks undersampling demonstrated reasonable performance overall, albeit slightly less effective than NearMiss.…”

Impact of Nature of Medical Data on Machine and Deep Learning for Imbalanced Datasets: Clinical Validity of SMOTE Is Questionable

“…Each of these approaches aims to mitigate the skewness inherent in imbalanced datasets, enhancing the model's ability to accurately identify and classify minority-class instances. Hence, the former involves resampling the training data to balance classes (e.g., undersampling, oversampling, or hybrid methods) [9][10][11][12][13], while the latter involves altering the classifier's learning process, known as algorithmic modifications (e.g., cost-sensitive learning, thresholding, or ensemble methods) [14,15]. Haixiang et al have demonstrated that in the domain of medical datasets, re-sampling-based ensemble classifiers are extensively utilized to tackle imbalances [4].…”