Χρήστου-Βαρσακέλης, ΔημήτριοςΚαπεταδημήτρη, ΓεωργίαΔιπλωματική εργασία--Πανεπιστήμιο Μακεδονίας, Θεσσαλονίκη, 2024.Made available in DSpace on 2024-07-19T08:19:51Z (GMT). No. of bitstreams: 2 license_rdf: 1031 bytes, checksum: 934f4ca17e109e0a05eaeaba504d7ce4 (MD5) KapetadimitriGeorgiaMsc2024.pdf: 6339334 bytes, checksum: 68ed2894f28e7c662386d6ed8964532b (MD5) Previous issue date: 2024-07-19Approved for entry into archive by Κυριακή Μπαλτά (balta@uom.gr) on 2024-07-19T08:19:51Z (GMT) No. of bitstreams: 2 license_rdf: 1031 bytes, checksum: 934f4ca17e109e0a05eaeaba504d7ce4 (MD5) KapetadimitriGeorgiaMsc2024.pdf: 6339334 bytes, checksum: 68ed2894f28e7c662386d6ed8964532b (MD5)Submitted by ΓΕΩΡΓΙΑ ΚΑΠΕΤΑΔΗΜΗΤΡΗ (aid23007@uom.edu.gr) on 2024-07-19T08:03:39Z No. of bitstreams: 2 license_rdf: 1031 bytes, checksum: 934f4ca17e109e0a05eaeaba504d7ce4 (MD5) KapetadimitriGeorgiaMsc2024.pdf: 6339334 bytes, checksum: 68ed2894f28e7c662386d6ed8964532b (MD5)Graph Neural Networks are widely employed for node classification in attributed networks. When it comes to fraud detection, however, GNNs can perform poorly, because a node’s features are typically computed based on its local neighborhood, and this allows fraudsters to "blend in" among legitimate users. In this thesis, GNNs and supervised contrastive learning are proposed for fraud detection on datasets where fraudsters mayuse intricate strategies to camouflage themselves within the network. We train our GNNs using novel structural features in addition to those typically used in similar studies. The proposed features are based on the empirical probability distributions of various graph structural attributes which are extracted from a given dataset. We also apply supervised contrastive learning, enhanced with synthetic samples for the minority class (i.e., the fraudsters). Under our approach, the classifying capability of the GNN(measured via F1-macro, AUC, Recall) is improved by boosting the representation power of the calculated embeddings that maximize the similarity between legitimate users while minimizing that between fraudsters and legitimate users. Numerical experiments on two real-world multi-relation graph datasets (Amazon and YelpChi) demonstrate the effectiveness of the proposed method, whose improvements over the state-of the-art were especially significant in the larger YelpChi dataset.http://dspace.lib.uom.gr/handle/2159/30975Πανεπιστήμιο ΜακεδονίαςΑναφορά Δημιουργού - Μη Εμπορική Χρήση - Παρόμοια Διανομή 4.0 Διεθνέςhttp://creativecommons.org/licenses/by-nc-sa/4.0/Contrastive LearningGraph Neural NetworksEnhancing fraud detection via gnns with synthetic fraud node generation and integrated structural featuresElectronic Thesis or DissertationText2024-07-19T08:19:51ZΤεχνητή νοημοσύνηArtificial Intelligenceisi - Computer Science, Artificial Intelligence covers resources that focus on research and techniques to create machines that attempt to efficiently reason, problem-solve, use knowledge representation, and perform analysis of contradictory or ambiguous information. This category includes resources on artificial intelligence technologies such as expert systems, fuzzy systems, natural language processing, speech recognition, pattern recognition, computer vision, decision-support systems, knowledge bases, and neural networks.Μεταπτυχιακή εργασίαMaster thesisΠανεπιστήμιο ΜακεδονίαςGreek Aggregator OpenArchives.gr | National Documentation Centre (EKT)