Minority report in fraud detection: classification of skewed data Export

@article{citeulike:3454032, abstract = {This paper proposes an innovative fraud detection method, built upon existing fraud detection research and Minority Report , to deal with the data mining problem of skewed data distributions. This method uses backpropagation (BP), together with naive Bayesian (NB) and C4.5 algorithms, on data partitions derived from minority oversampling with replacement. Its originality lies in the use of a single meta-classifier (stacking) to choose the best base classifiers, and then combine these base classifiers' predictions (bagging) to improve cost savings (stacking-bagging). Results from a publicly available automobile insurance fraud detection data set demonstrate that stacking-bagging performs slightly better than the best performing bagged algorithm, C4.5, and its best classifier, C4.5 (2), in terms of cost savings. Stacking-bagging also outperforms the common technique used in industry (BP without both sampling and partitioning). Subsequently, this paper compares the new fraud detection method (meta-learning approach) against C4.5 trained using undersampling, oversampling, and SMOTEing without partitioning (sampling approach). Results show that, given a fixed decision threshold and cost matrix, the partitioning and multiple algorithms approach achieves marginally higher cost savings than varying the entire training data set with different class distributions. The most interesting find is confirming that the combination of classifiers to produce the best cost savings has its contributions from all three algorithms.}, address = {New York, NY, USA}, author = {Phua, Clifton and Alahakoon, Damminda and Lee, Vincent}, citeulike-article-id = {3454032}, citeulike-linkout-0 = {http://portal.acm.org/citation.cfm?id=1007730.1007738}, citeulike-linkout-1 = {http://dx.doi.org/10.1145/1007730.1007738}, doi = {10.1145/1007730.1007738}, issn = {1931-0145}, journal = {SIGKDD Explor. Newsl.}, keywords = {classification, fraud, skew}, number = {1}, pages = {50--59}, posted-at = {2008-12-04 23:20:09}, priority = {2}, publisher = {ACM}, title = {Minority report in fraud detection: classification of skewed data}, url = {http://dx.doi.org/10.1145/1007730.1007738}, volume = {6}, year = {2004} }

TY - JOUR ID - citeulike:3454032 L3 - citeulike-article-id:3454032 N2 - This paper proposes an innovative fraud detection method, built upon existing fraud detection research and Minority Report , to deal with the data mining problem of skewed data distributions. This method uses backpropagation (BP), together with naive Bayesian (NB) and C4.5 algorithms, on data partitions derived from minority oversampling with replacement. Its originality lies in the use of a single meta-classifier (stacking) to choose the best base classifiers, and then combine these base classifiers' predictions (bagging) to improve cost savings (stacking-bagging). Results from a publicly available automobile insurance fraud detection data set demonstrate that stacking-bagging performs slightly better than the best performing bagged algorithm, C4.5, and its best classifier, C4.5 (2), in terms of cost savings. Stacking-bagging also outperforms the common technique used in industry (BP without both sampling and partitioning). Subsequently, this paper compares the new fraud detection method (meta-learning approach) against C4.5 trained using undersampling, oversampling, and SMOTEing without partitioning (sampling approach). Results show that, given a fixed decision threshold and cost matrix, the partitioning and multiple algorithms approach achieves marginally higher cost savings than varying the entire training data set with different class distributions. The most interesting find is confirming that the combination of classifiers to produce the best cost savings has its contributions from all three algorithms. IS - 1 JF - SIGKDD Explor. Newsl. SN - 1931-0145 AD - New York, NY, USA EP - 59 TI - Minority report in fraud detection: classification of skewed data PB - ACM VL - 6 SP - 50 KW - classification KW - fraud KW - skew AU - Phua, Clifton AU - Alahakoon, Damminda AU - Lee, Vincent PY - 2004/// UR - http://dx.doi.org/10.1145/1007730.1007738 ER -