A Suspicious Company Detection System Based on Machine Learning with a Bibliometric Analysis for Accounting and Auditing

Authors : Yunus Korkmaz, Sadık Serçek, Mukaddes Korkmaz

Pages : 164-174

Doi:10.46460/ijiea.1789267

View : 61 | Download : 172

Publication Date : 2025-12-29

Article Type : Research Paper

Abstract :In accounting, auditing involves reviewing the financial statements and records of an organization by independent auditors in order to guarantee their precision and comply with relevant statutes and rules. Audits are typically conducted by external or internal auditors who review financial statements, assess internal controls, and verify the precision of financial data. The aim of audit is to provide confidence to stakeholders that the pecuniary data presented by the organization is reliable and trustworthy. In this work, role of the artificial intelligence (AI) was examined with a comprehensive literature review for auditing in accounting area. In technical aspect of this work, we built a fraudulent company detection system based on machine learning (ML) classification algorithms like Decision Tree (DT), Bagged Tree, Ensemble KNN (K-Nearest Neighbors), Linear Support Vector Machines (SVM) etc. Decision Tree and Bagged Tree reached AUC value of 1 which means that they are the perfect classifiers. The AUC values of 0.99, 0.99, 0.9998 and 0.9963 were obtained for Linear SVM, Logistic Regression, Subspace KNN and Naïve Bayes, respectively. The result proved that any machine learning based solution can help auditors to easily have an idea about the fraudulent companies before on-site auditing.
Keywords : Machine Learning, Suspicious Company, Decision Tree, Bibliometric Analysis, Accounting and Auditing

ORIGINAL ARTICLE URL