Uncertainty-aware credit card fraud detection using deep learning

Habibpour, Maryam; Gharoun, Hassan; Mehdipour, Mohammadreza; Tajally, AmirReza; Asgharnezhad, Hamzeh; Jokandan, Afshar Shamsi; Khosravi, Abbas; Shafie‐khah, Miadreza; Nahavandi, Saeid; Catalão, João P.S.

doi:10.1016/j.engappai.2023.106248

Cited by 19 publications

(2 citation statements)

References 24 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In contrast, ML techniques are employed for precise decisions -One well-known approach for classifying online transactions is to differentiate fraud with typical labels in the training set, often called supervised learning [3]. Common approaches in this subject comprise logistic regression [4], k-nearest neighbors [5], support vector machines [6], and decision trees [7]. This strategy uses labeled past transactions to create a predictive fraud algorithm that predicts the chance of every new transaction being fraudulent.…”

Section: Introductionmentioning

confidence: 99%

Real-Time Online Banking Fraud Detection Model by Unsupervised Learning Fusion

Abbassi,

El Mendili,

Gahi

2024

HighTech. Innov. J.

View full text Add to dashboard Cite

Digital trades and payments are becoming increasingly popular, as they typically entail monetary transactions. This not only makes electronic transactions more convenient for the end customer, but it also raises the likelihood of fraud. An adequate fraud detection system with a cutting-edge model is critical to minimizing fraud costs. Identifying fraud at the ideal time entails establishing and setting up ubiquitous systems to consume and analyze massive amounts of streaming data. Recent advances in data analytics methods and introducing open-source technology for big data storage and processing opened new options for detecting fraud. This study aims to tackle this critical issue by providing a newly real-time e-transaction fraud detection schema that consolidates the advantages of both unsupervised learners, including autoencoder and extended isolation forests, with cutting-edge big data gadgets such as Spark streaming and sparkling water. It addresses the shortage of non-fraudulent instances and handles the excessive dimension of the set of features. On two real-world transactional datasets, we assess our suggested technique. Compared with other current fraud identification systems, our methodology delivers an elevated accuracy yield of 99%. Furthermore, it outperforms state-of-the-art approaches in reliably identifying fraudulent samples. Doi: 10.28991/HIJ-2024-05-01-014 Full Text: PDF

show abstract

Section: Introductionmentioning

confidence: 99%

Real-Time Online Banking Fraud Detection Model by Unsupervised Learning Fusion

Abbassi,

El Mendili,

Gahi

2024

HighTech. Innov. J.

View full text Add to dashboard Cite

show abstract

“…By using CNNs, different VGG models, and RNNs, it is possible to improve the capabilities of both the local radiologist and the outsourced team regarding the precise diagnosis of complex cases. It may further ease caseloads for radiologists, enhance efficiency, and ensure the required accuracy of the diagnosis, in turn improving patient outcomes [16], [17].…”

Section: Introductionmentioning

confidence: 99%

A Comparative analysis of Advanced Machine Learning Techniques for Enhancing Brain Tumor Detection

G. Nagaraju

2024

jes

View full text Add to dashboard Cite

Detection and classification of brain tumors are prominent tasks in neuroimaging, as it ensures accurate and timely diagnosis for effective treatment planning. This study explores the utility of machine learning models, such as convolutional neural networks, VGG 16 and 19 architectures, and recurrent neural networks, in improving the detection of brain tumor anomaly in both CT and MRI. A total of 3400 images were integrated and collected from multiple sources and prepared for usage by conducting meticulous preprocessing and feature extraction. Following the preparation process, the models were trained using stratified dataset split with a 70/30 ratio. The testing results indicated that the VGG 16 and VGG 19 architectures yielded the highest performance results, as they produced the optimal precision, recall, and F1-score values that reached up to 96.7%, 97.2%, and 96.5%, respectively, along with the highest AUC-ROC scores. In comparison, the CNN and RNN models presented lower performance results in each measured metric. The superiority of the VGG architectures strengthes the idea that the complexity and the capacity of the model to memorize and retain the imaging features is crucial for the accurate detection of the tumors. As such, the improved results can be helpful for healthcare professionals, as they provide powerful tools to ensure precise detection and characteristics of brain tumors. Nevertheless, further research is required to validate the results with a larger dataset, and the adherence of the study to the black-box nature of the model poses limitations on the interpretation. In any case, the study supplements the growing body of research in medical imaging and contributes to the prevention and management of brain tumors.

show abstract