International Conference on Computer and Knowledge Engineering

Home / 15th International Conference on Computer and Knowledge Engineering

Semi-Supervised Supply Chain Fraud Detection with Unsupervised Pre-Filtering

Authors :

Fatemeh Moradi¹ Mehran Tarif² Mohammadhossein Homaei³

1- Faculty of Engineering Isfahan (Khorasgan) Branch, Islamic Azad University Isfahan, Iran 2- Department of Computer Science University of Verona Verona, Italy 3- Media Engineering Group University of Extremadura C´aceres, Spain

Keywords :

Supply chain fraud detection, , , , ,،Isolation Forest،Self-training SVM،Semi-supervised learning،Anomaly detection،Class imbalance

Abstract :

Detecting fraud in modern supply chains is a grow-ing challenge, driven by the complexity of global networksand the scarcity of labeled data. Traditional detection methodsoften struggle with class imbalance and limited supervision,reducing their effectiveness in real-world applications. This paperproposes a novel two-phase learning framework to address thesechallenges. In the first phase, the Isolation Forest algorithmperforms unsupervised anomaly detection to identify potentialfraud cases and reduce the volume of data requiring furtheranalysis. In the second phase, a self-training Support VectorMachine (SVM) refines the predictions using both labeled andhigh-confidence pseudo-labeled samples, enabling robust semi-supervised learning. The proposed method is evaluated on theDataCo Smart Supply Chain Dataset, a comprehensive real-worldsupply chain dataset with fraud indicators. It achieves an F1-score of 0.817 while maintaining a false positive rate below 3.0%.These results demonstrate the effectiveness and efficiency of com-bining unsupervised pre-filtering with semi-supervised refinementfor supply chain fraud detection under real-world constraints,though we acknowledge limitations regarding concept drift andthe need for comparison with deep learning approaches.