Data mining for better fraud detection

This research objective focuses on data mining, which is quickly becoming a popular artificial intelligence technique used by tax administrations for better fraud detection. However, many challenges still exist. One of the main challenges is explaining the predicition made by complex algorithms. Within this research objective, we explore the opportunities of tackling this challenge of explainable AI by making use of counterfactual explanations. 

​Fraud detection with behavioural data

• researcher Dieter Brughmans 
• supervisor prof. David Martens 
• research is funded by the University of Antwerp
• read more on this topic on the website of Antwerp Tax Academy​​

Explainable Artificial Intelligence  

• researcher Raphael Mazzine  
• supervisor David Martens 
• research is funded by the Flemish Government​
• read more on this topic on the website of Antwerp Tax Academy 

​​​​​​​

2022

• Martens David, ata science ethics : concepts, techniques, and cautionary tale, Oxford University Press 2022, 272 p.
• Goethals Sofie, Martens David, Calders Toon, PreCoF: counterfactual explanations for fairness, Research Square 2022, 31 p.
  
• Vermeire Tom, Brughmans Dieter, Goethals Sofie, Mazzine Barbossa de Oliveira Raphael, Martens David, Explainable image classification with evidence counterfactual, Pattern Analysis and Applications 2022, 315-335 ​
  
• Greene Travis, Martens David, Shmueli Galid, Barriers for academic data science research in the new realm of behavior modification by digital platforms, Nature Machine Intelligence 2022, p. 323-330

2021

• Brughmans Dieter en Martens David, NICE : an algorithm for nearest instance counterfactual explanations, ArXiv 2021, 16 p. 
• De Raedt Sylvie, Brughmans Dieter en Martens David: Waarom krijg ik fiscale controle? Naar meer transparantie bij de geautomatiseerde besluitvorming bij de fiscale overheid, Tijdschrift voor Fiscaal Recht, 2021, 607-612 ​​
• Vermeire Tom, Laugel Thibault, Renard Xavier, Martens David, Detyniecki Marcin, How to choose an explainability method? Towards a methodical implementation of XAI in practice , ArXiv, 2021, 16 p.
• Ramon Yanou, Vermeire Tom, Toubia Olivier, Martens David, Evgeniou Theodoros, Understanding consumer preferences for explanations generated by XAI algorithms , Arvix 2021, 18 p.
• Mazzine Raphael, Martens David, A framework and benchmarking study for counterfactual generating methods on tabular data, ArXiv 2021, 33 p. and Applied Sciences 2021, 28 p. 
• Greene Travis, Martens David, Shmueli Ghalit, Barriers for academic data science research in the new realm of behavior modification by digital platforms, ssrn, 2021, 1-15
• Ramon Yanou, Martens David, Evgeniou Theodoros, Praet Stiene, Can metafeatures help improve explanations of prediction models when using behavioral and textual data?, Machine Leaning, Dordrecht, 2021, 40 p.
  
• Ramon Yanou, Farrokhnia R.A.,Matz Sandra C., Martens David, Explainable AI for psychological profiling from behavioral data : an application to big five personality predictions from financial transaction records,  ArXiv 2021, 24 p. 

2020

• Vermeire, Tom and Martens, David, "Explainable Image Classification with Evidence Counterfactual", working paper April 16, 2020
  
• ​Vanhoeyveld Jellis, Martens David, Peeters Bruno, "Value-added tax fraud detection with scalable anomaly detection techniques", in Applied soft computing, 86 (2020), 1-2°​​
  
• Ramon Yanou, Martens David, Provost Foster, Evgeniou Theodoros, A comparison of instance-level counterfactual explanation algorithms for behavioral and textual data : SEDC, LIME-C and SHAP-, Advances in Data Analysis and Classification, 2020, 801-819

2019

• Vanhoeyveld, Jellis, Martens, David, Peeters, Bruno, "Customs fraud detection : assessing the value of behavioural and high-cardinality data under the imbalanced learning issue", in Pattern analysis and applications, London - New york,  Springer, 2019, 1433-7541 (21 p.)
  
•  Vanhoeyveld, Jellis, "Data mining for tax fraud detection", University of Antwerp, Faculty of Business and Economics, 2019, 281 p. (doctoral thesis)

Previous Research

2018

• Vanhoeyveld, Jellis, Martens, David, "Imbalanced classification in sparse and large behaviour datasets", in Data mining and knowledge discovery, Boston, Mass., 2018, 32:1 (2018), p. 25-82
• Vanhoeyveld, Jelllis, Martens, David, "Towards a scalable anomaly detection with pseudo-optimal hyperparameters", Research paper / University of Antwerp, Faculty of Business and Economics ; 2018-012, 59 p.

2016

•  Van Hoeyveld, Jellis, Martens, David, Peeters, Bruno, "Datamining voor fraudedetectie",  in Cahiers politiestudies / Centrum voor Politiestudies,  Gent, 2016, 39:2(2016), p. 167-211
•  Martens, David, De Cnudde, Sofie, Moeyersoms, Julie, Praet, Stiene, Stankova, Marija, Tobback, Ellen, Vanhoeyveld, Jellis, "Big data in banking", in Bank- en financiewezen / Belgisch Financieel Forum. Studiecentrum voor het Financiewezen; Forum financier belge. Centre d'études financières, Brussel, 2016, p. 117-122
•  Vanhoeyveld, Jellis, Martens, David, Peeters, Bruno, "Datamining voor fraudedetectie", in Criminele organisaties en organisatiecriminaliteit, Serie Cahiers politiestudies / Centrum voor Politiestudies, Antwerpen, Maklu, 2016, p. 167-211