<![CDATA[Discrepancies in payment systems, such as overpayments, underpayments, and recording errors, have the potential to cause financial losses and reduce institutional accountability. This study aims to detect anomalies in the payment data of Lecturer Professional Allowances (TPD) and Distinguished Professor Honoraria (TKGB) at LLDIKTI Region IV using the Isolation Forest algorithm, and to evaluate the associated financial risks through the COSO ERM framework. The data analyzed were derived from historical SPTJM Online records. The results show that the algorithm successfully identified 150 anomalies in salary data and 144 in payment data, with significant deviation scores. t-SNE visualization revealed a clear separation between normal and anomalous data, while the chi square test indicated that the anomalies were systemic in nature. The COSO ERM evaluation highlighted the highest compliance in risk identification, although weaknesses were found in data integration and reporting systems. This integrative approach proves effective in detecting anomalies and strengthening financial oversight in higher education institutions.]]>
Copyrights © 2025
