Opinion Detection of Public Sector Financial Statements Using K-Nearest Neighbors

Ahmad Dwi Arianto
Achmad Affandi
Supeno Mardi Susiki Nugroho

Proceeding of the Electrical Engineering Computer Science and Informatics
Vol 4: EECSI 2017

The identification of ethical violations committedby the auditor is very difficult to do. Artificial intelligence offersanomaly detection as an alternative method for detecting theopinion anomaly which can be an early indicator of the opiniontrading occurrence. This paper proposes the use of originalfeatures from public sector rather than the use of modifiedfeatures from the private sector to be applied in opinion detectionin public sector. By using 60% Holdout validation, 1-NNclassification showed that original featured from the public sectoroutperformed the modified featured from the private sector by5.82% through 13.10% under F-Measure Criterion and by4.22% through 9.56% under AUC criterion.