FINANCIAL DATA ANOMALY DETECTION METHOD BASED ONRESIDUAL EXPLANATION FOR GOVERNMENT OFFICEOPERATIONAL EXPENDITURE

Abstract

Ensuring fiscal responsibility and transparency in government offices requirescareful control of operating expenditures. However, due to the intricacy and regularity offinancial transactions, this work can be difficult, making standard monitoring techniquesinefficient and time-consuming. This paper introduces an anomaly detection method forfinancial data in government office operational expenditures, termed "ResidualExplanation". This approach uses sophisticated machine learning techniques to discoveranomalous transaction patterns that can point to fraud by analyzing residuals, or thedisparities between observed and predicted transaction values. Our method makes use of aRandom Forest Regressor model, which is especially well-suited to managing the highdimensionality and non-linear correlations seen in financial datasets. This study focuses onanomaly detection within operational expenses of account 525112 at one of institutions inIndonesia. The results indicate that the optimal model for detecting anomalies operates witha thresholds value of 99,6%. Future improvements to this model could allow its integrationof our approach into existing financial systems could enable real-time anomaly detection,which is paramount for preventing fraud and enhancing the financial governance ofgovernment expenditures.