Nos derniers articles

This document introduces a novel two-step methodology for money laundering detection that significantly improves upon existing rule-based and traditional machine learning methods. The first step involves representation learning using a transformer neural network, which analyzes complex financial time series data without requiring labels through contrastive learning. This self-supervised pre-training helps the model understand the inherent patterns in transactions. The second step then leverages these learned representations within a two-threshold classification procedure, calibrated by the Benjamini-Hochberg (BH) procedure, to control the false positive rate while accurately identifying both fraudulent and non-fraudulent accounts, addressing the significant class imbalance in money laundering datasets. Experimental results on real-world, anonymized financial data demonstrate that this transformer-based approach outperforms other models in detecting fraudulent activities.

This study develops a machine learning framework to identify high-risk enterprise financial reports, comparing Support Vector Machine, Random Forest, and K-Nearest Neighbors models. Using 2020–2025 audit data from the Big Four firms, Random Forest showed the highest performance (F1-score: 0.9012), excelling in detecting fraud and compliance issues. While KNN struggled with high-dimensional data, SVM performed well but was computationally intensive. The study highlights the potential of machine learning in auditing but notes limitations, including reliance on structured data and exclusion of external economic factors.

This study proposes an attention-based ensemble model for detecting credit card fraud, integrating classifiers' predictions using two aggregation operators (DOWA and IOWA). The model, which selects key features via a bootstrap forest, achieves 99.95% accuracy and a perfect AUC of 1, demonstrating the effectiveness of AI in fraud detection.

The paper explores credit card fraud detection (CCFD) using machine learning, reviewing various algorithms like K-nearest neighbors, decision trees, random forests, and XGBoost. It compares their performance, highlighting Random Forest as the most accurate. The study addresses challenges like imbalanced datasets, data quality, and evolving fraud tactics.

The research examines the impact of traditional verification methods on customer satisfaction and operational efficiency in commercial banks. It suggests adopting digital solutions such as biometric authentication and machine learning algorithms to streamline verification, prevent fraud, enhance security, and improve customer engagement. The study aims to provide recommendations for innovative methods aligning with digital transformation and customer expectations.

Online transaction fraud poses significant challenges to businesses and consumers, with rule-based systems struggling to keep up. Machine learning, particularly personalized PageRank (PPR), offers promise by analyzing account relationships. Results show PPR enhances fraud detection models, providing valuable insights and stable features across datasets, improving predictive power.