Predicting fraud is challenging due to inherent issues in the fraud data structure, since the crimes are committed through trickery or deceit with an ever-present moving target of changing modus operandi to circumvent human and system controls. As a national security challenge, criminals continually exploit the electronic financial system to defraud consumers and businesses by finding weaknesses in the system, including in audit controls. This study uses stacked generalization using meta or super learners for improving the performance of algorithms in step one (minimizing the algorithm error rate to reduce its bias in the learning set) and then in step two the results are input into the meta learner with its stacked blended output (with the weakest algorithms learning better). A fundamental key to fraud data is that it is inherently not systematic, and an optimal resampling methodology has yet not been identified. Building a test harness, for all permutations of algorithm sample set pairs, demonstrates that the complex, intrinsic data structures are all thoroughly tested. A comparative analysis on fraud data that applies stacked generalizations provides useful insight to find the optimal mathematical formula for imbalanced fraud data sets necessary to improve upon fraud detection for national security.
ACFE. Professional training promoting the development and implementation of sound anti-money laundering policies and procedures, https://www.acfe.com/report-to-the-nations/2018 (2018, accessed March–May 2019).
2.
MillsC.Predictive analytics in fraud and AML. J Fin Complian2017; 1: 17–26.
3.
SweckerCBrothersB. Crime rings have grown to a prevalent and formidable threat, association of certified anti-money launderer specialists (ACAMS) webinar training, https://lmsacams.becker.com/media/813985 (2015, accessed March–May 2019).
4.
SweckerCBrothersB. AML: a new approach to fight complex money laundering schemes, association of certified anti-money launderer specialists (ACAMS) webinar training, https://lmsacams.becker.com/media/816463 (2015, accessed March–May 2019).
EuropolS. Serious and organized crime threat assessment, identifying the priorities in the fight against major crime, https://www.europol.europa.eu/socta-report (2017, accessed March–May 2019).
YusofRKasmiranKAMustaphaA, et al. Techniques for handling imbalanced datasets when producing classifier models. J Theor Appl Inform Technol2017; 95: 1425–1440.
15.
JapkowiczN. Learning from imbalanced data sets: a comparison of various strategies. AAAI workshop on learning from imbalanced data sets2000; 68: 10–15.
16.
KotsiantisSKanellopoulosDPintelasP, et al. Handling imbalanced datasets: a review. GESTS Int Trans Comput Sci Eng2006; 30: 25–36.
17.
ChawlaNV. Data mining for imbalanced datasets: an overview. In: MaimonORokachL (eds) Data mining and knowledge discovery handbook. Boston, MA: Springer, 2005, pp.853–867.
MaaloufMTrafalisTB.Rare events and imbalanced datasets: an overview. Int J Data Mining Model Manag2011; 3: 375–388.
20.
ChenZTeohENNazirA, et al. Machine learning techniques for anti-money laundering (AML) solutions in suspicious transaction detection: a review. Knowl Inform Syst2018; 57: 245–285.
21.
WeissGMProvostF.Learning when training data are costly: the effect of class distribution on tree induction. J Artif Intell Res2003; 19: 315–354.
22.
ChawlaNV. C4. 5 and imbalanced data sets: investigating the effect of sampling method, probabilistic estimate, and decision tree structure. In: proceedings of the ICML, vol. 3, 2003, p.66.
LingCXLiC. Data mining for direct marketing: problems and solutions. In: Kdd, volume 98. pp.73–79.
25.
JapkowiczN. The class imbalance problem: significance and strategies. In: proceedings of the international conference on artificial intelligence (IC-AI 2000), Las Vegas, NV, 26 June 2000.
26.
McCarthyKZabarBWeissG. Does cost-sensitive learning beat sampling for classifying rare classes? In: proceedings of the 1st international workshop on utility-based data mining, pp.69–77. New York: Association for Computing Machinery.
27.
WeissGMMcCarthyKZabarB.Cost-sensitive learning vs. Sampling: which is best for handling unbalanced classes with unequal error costs?Dmin2007; 7: 24.
28.
RajeswariK.Feature selection by mining optimized association rules based on a priori algorithm. Int J Comput Appl2015; 119.
29.
ZhengZWuXSrihariR.Feature selection for text categorization on imbalanced data. ACM Sigkdd Explorations Newsletter2004; 6: 80–89.
TingKMWittenIH.Issues in stacked generalization. J Artif Intell Res1999; 10: 271–289.
32.
ZhangYWangD. A cost-sensitive ensemble method for class-imbalanced datasets. In: Abstract and applied analysis, volume 2013. Hindawi.
33.
SesmeroMPLedezmaAISanchisA.Generating ensembles of heterogeneous classifiers using stacked generalization. Wiley Interdisc Rev Data Mining Knowl Discov2015; 5: 21–34.
34.
NaimiABalzerL.Stacked generalization: an introduction to super learning. Eur J Epidemiol2018; 33: 459–464.
35.
WongMLSengKWongPK.Cost-sensitive ensemble of stacked denoising autoencoders for class imbalance problems in business domain. Expert Syst Appl2020; 141.
36.
WinstonPOrtizL.6.034f boosting notes, 2009. Patrick Winston and Luis Ortiz Draft of November 17, 2009, MIT - Massachusetts Institute of Technology.
37.
WangJYouJLiQ, et al. Extract minimum positive and maximum negative features for imbalanced binary classification. Patt Recognit2012; 45: 1136–1145.
38.
TingKMWittenIH.Stacked generalization: when does it work? In: international joint conferences on artificial intelligence, University of Waik, 1997.
39.
BontempiGLenaertsT. Machine Learning Group - ULB Machine Learning Group (MLG), a research unit of the Computer Science Department of the Faculty of Sciences. Creditcard.csv dataset, https://www.kaggle.com/mlg-ulb (2004, accessed March–May 2019).
40.
HuLYHuangMWKeSW, et al. The distance function effect on k-nearest neighbor classification for medical datasets. SpringerPlus2016; 5: 1–9.
TomekI.Two modifications of CNN. IEEE Trans Syst Man Cybern1976; 6: 769–772.
43.
HanHWangWYMaoBH. Borderline-smote: a new over-sampling method in imbalanced data sets learning. In: international conference on intelligent computing, pp.878–887. Springer.
44.
HeHBaiYGarciaEA, et al. ADASYN: adaptive synthetic sampling approach for imbalanced learning. In: 2008 IEEE international joint conference on neural networks (IEEE world congress on computational intelligence), 2008, pp.1322–1328. IEEE.
45.
QuinlanJR. Decision trees as probabilistic classifiers. In: proceedings of the fourth international workshop on machine learning, 1987, pp.31–37. Elsevier.
46.
SharmaSAgrawalJSharmaS.Classification through machine learning technique: C4. 5 algorithm based on various entropies. Int J Comput Appl2013; 82.
47.
JamesGWittenDHastieT, et al. An introduction to statistical learning, volume 112. New York: Springer, 2013.
48.
NgA. Deep Learning School on September 24/25, 2016, Video file. Nuts and bolts of applying deep learning, https://www.youtube.com/watch?v=F1ka6a13S9I (2016, accessed March–May 2019).
49.
TurneyPD.Types of cost in inductive concept learning. arXiv preprint cs/02120342002.
50.
ShilbayehSA, et al. Cost sensitive meta-learning. PhD Thesis, University of Salford, 2015.
51.
HandDJWhitrowCAdamsNM, et al. Performance criteria for plastic card fraud detection tools. J Oper Res Soc2008; 59: 956–962.
52.
BahnsenACVillegasSAouadaD, et al. Fraud detection by stacking cost-sensitive decision trees. Data Science for Cyber-Security. World Scientific, 2017.
53.
LomaxSVaderaS.A survey of cost-sensitive decision tree induction algorithms. ACM Comput Surv2013; 45: 1–35.
54.
ShannonCE.A mathematical theory of communication. Bell Syst Techn J1948; 27: 379–423.
55.
LomaxSE, et al. Cost-sensitive decision tree learning using a multi-armed bandit framework. PhD Thesis, University of Salford, 2013.
56.
EfronB.Computers and the theory of statistics: thinking the unthinkable. SIAM Rev1979; 21: 460–480.
57.
KrausA.Recent methods from statistics and machine learning for credit scoring. PhD Thesis, LMU, 2014.
YinQYZhangJSZhangCX, et al. An empirical study on the performance of cost-sensitive boosting algorithms with different levels of class imbalance. Math Probl Eng2013; 2013.
60.
NatekinAKnollA.Gradient boosting machines, a tutorial. Frontier Neurorobot2013; 7: 21.
61.
LiuXYWuJZhouZH.Exploratory undersampling for class-imbalance learning. IEEE Trans Syst Man Cybern B2008; 39: 539–550.
62.
SeiffertCKhoshgoftaarTMVan HulseJ, et al. Rusboost: a hybrid approach to alleviating class imbalance. IEEE Trans Syst Man Cybern A Syst Hum2009; 40: 185–197.
63.
YanJHanS.Classifying imbalanced data sets by a novel re-sample and cost-sensitive stacked generalization method. Math Probl Eng2018; 2018.
64.
FawcettTProvostF.Adaptive fraud detection. Data Mining Knowl Discov1997; 1: 291–316.
65.
FawcettT.An introduction to roc analysis. Patt Recognit Lett2006; 27: 861–874.
HeHGarciaEA.Learning from imbalanced data. IEEE Trans Knowl Data Eng2009; 21: 1263–1284.
69.
JoshiMVAgarwalRCKumarV. Predicting rare classes: can boosting make any weak learner strong? In: proceedings of the eighth ACM SIGKDD international conference on knowledge discovery and data mining, 2002, pp.297–306.
