METHODS OF DETECTING ANOMALIES IN BUSINESS DATA USING MACHINE LEARNING: A COMPARATIVE REVIEW OF MODELS AND PRACTICAL CASES

Abstract

This article presents a comparative review of methods for detecting anomalies in business data using machine learning. Statistical, cluster, density, neural network , and hybrid approaches, along with their advantages and limitations, are discussed. Particular attention is paid to practical cases: fraudulent transaction detection, supplier auditing, data quality control, and operational process monitoring.