Electronic Thesis/Dissertation
 

Enhancing Detections of Occupational Fraud in Purchases within the Manufacturing Industry Using Machine Learning

Open Access Deposited

XGBoost, SAINT, and LinearBoost Classifier. Findings show that with feature engineering and multiple instance learning (MIL) each of the three algorithms can achieve equal performance on the dataset, outperforming the accuracy of existing models.

Deficient monitoring of procurement processes in manufacturing organizations limits their ability to detect occupational fraud, contributing to a 5% revenue loss per year. The application of machine learning has the potential to detect acts of fraud earlier and more comprehensively than current human and software detection systems. This research uses machine learning to develop an efficient fraud detection and classification model, trained on multiple perspectives, and which is resilient to variations in an organization’s processes. Critical features detected in transaction logs of an SAP ERP dataset are identified and examined as through a microscope using algorithm visualization, scenario-to-text generation, and feature explanation tools. Supporting the need to balance performance and environmental impact, three machine-learning algorithms differing in core levels of complexity are compared in terms of accuracy

Author Language Date created Type of Work License
  • All rights reserved
Rights statement GW Unit Degree Advisor Committee Member(s) Persistent URL

Notice to Authors

If you are the author of this work and you have any questions about the information on this page, please use the Contact form to get in touch with us.

Thumbnail Title Date Uploaded Visibility Actions
Preview of Reichenthal_gwu_0075A_17710.pdf Reichenthal_gwu_0075A_17710.pdf 2025-12-15 Open Access