ANOMALY DETECTION IN PAYROLL TRANSACTIONS USING EXPLAINABLE MACHINE LEARNING MODELS

Abstract

Payroll transaction management is a critical financial process in enterprises where accuracy, transparency, and fraud prevention are essential for maintaining organizational trust and regulatory compliance. Traditional payroll auditing approaches rely on manual verification, predefined rules, and periodic reviews, which are often ineffective in detecting complex and hidden transaction anomalies. This paper proposes an Anomaly Detection Framework in Payroll Transactions Using Explainable Machine Learning Models by integrating machine learning, anomaly detection algorithms, explainable artificial intelligence (XAI), financial analytics, automated data pipelines, and enterprise payroll systems. The proposed framework analyzes payroll transaction patterns, employee compensation records, historical payment behavior, and financial attributes to identify suspicious activities and irregularities. Explainable models provide transparent insights into detected anomalies and support effective decision-making by payroll administrators and auditors. Experimental evaluation demonstrates improvements in anomaly detection accuracy, interpretability, fraud identification capability, audit efficiency, and payroll reliability. The proposed framework provides a scalable and trustworthy solution for intelligent payroll monitoring in modern enterprises. Keywords— Payroll Analytics, Anomaly Detection, Explainable AI, Machine Learning, Financial Fraud Detection, Payroll Transactions, Intelligent Automation, Enterprise Analytics.