Research on financial data anomaly detection and intelligent auditing system combined with machine learning

With the increasing complexity and informatization of enterprise financial activities, the traditional auditing methods have gradually exposed the problems of low efficiency, poor accuracy, and heavy manual reliance in the face of large-scale, high-dimensional, and diversified financial data. In order to improve the intelligence and automation level of financial anomaly detection, this paper proposes a research framework of financial data anomaly detection and intelligent auditing system combined with machine learning technology. By constructing an analysis process that includes data preprocessing, feature engineering, anomaly detection modeling and system integration modules, this paper designs an efficient audit support system that realizes rapid identification and accurate positioning of abnormal transactions, financial fraud and other behaviors. The empirical results show that the system has good detection performance and scalability when dealing with real enterprise data, which is significantly better than the traditional rule-driven auditing method. This study provides theoretical support and practical path for the development of intelligent auditing technology, and also provides an important technical guarantee for enterprise risk control and financial compliance management.