FINANCIAL FRAUD DETECTION USING VALUE AT RISK WITH MACHINE LEARNING IN SKEWED DATA

Abstract

Financial fraud has emerged as a significant challenge in modern financial systems, causing substantial economic losses and undermining the integrity of financial transactions. Detecting fraudulent activities is particularly difficult due to the highly imbalanced and skewed nature of financial datasets, where fraudulent transactions constitute only a small fraction of the overall data. Traditional fraud detection approaches often struggle to accurately identify rare fraudulent events while maintaining low false alarm rates. This paper presents a financial fraud detection framework using Value at Risk (VaR) and machine learning techniques in skewed data environments. The proposed approach combines risk assessment through Value at Risk analysis with advanced machine learning algorithms to identify suspicious financial activities and estimate potential financial losses. Data preprocessing and imbalance handling techniques are employed to address skewed class distributions and improve model performance. Machine learning models analyze transaction patterns, behavioral characteristics, and financial risk indicators to distinguish between legitimate and fraudulent activities. Experimental results demonstrate that the integration of VaR with machine learning enhances fraud detection accuracy, improves risk prediction capabilities, and supports proactive financial risk management. The proposed framework provides an effective and scalable solution for detecting financial fraud in complex and highly imbalanced financial datasets.