Роскошенко В.В.аспирант, магистр экономики, экономический факультет, МГУ имени М.В. Ломоносова, Москва, Российская Федерация roskoshenkoeco@mail.ru ORCID id: отсутствует SPIN-код: отсутствует
Предмет. Проблема несбалансированности классов в выборочных данных при моделировании дефолта кредитного требования, подходы к предварительной обработке данных, позволяющие преодолеть дисбаланс классов. Имеющиеся исследования по сопоставлению таких подходов выполнены либо в отношении небольшого числа методов, либо на специфических данных из отдельных областей деятельности. Ранее в литературе не рассмотрены подходы на основе сочетания методов предварительной обработки данных с ансамблевым решением (стэкингом). Цели. Произвести поиск оптимального варианта по преодолению проблемы несбалансированности классов среди каждой из групп подходов для банковских данных о кредитовании физических лиц. Методология. Использованы математическое моделирование, статистический анализ и контент-анализ источников. Результаты. Показано, что подход EditedNearestNeighbours, будучи довольно сложным с вычислительной точки зрения, оказался оптимальным. В его основе — удаление представителей доминирующего класса, плохо удовлетворяющих своему окружению, которое определяется посредством кластеризации. Среди сочетаний подходов предварительной обработки данных и стэкинга оптимальным оказался вариант с RandomOverSampler. Последний предполагает увеличение доли миноритарного класса случайным образом и является одним из наиболее простых. Область применения. Результаты могут быть использованы в кредитном скоринге и в любом статистическом моделировании, где требуется бинарная классификация. Выводы. Осуществлено исчерпывающее сопоставление подходов по преодолению проблемы несбалансированности классов в выборочных данных. Были определены оптимальный подход среди подходов предварительной обработки данных, а также оптимальное сочетание подхода предварительной обработки данных с ансамблевым решением.
Sun Y., Wong A.C., Kamel M.S. Classification of imbalanced data: A review. International Journal of Pattern Recognition and Artificial Intelligence, 2009, vol. 23, no. 4, pp. 687–719. URL: Link
García V., Mollineda R., Sánchez J. On the k-NN performance in a challenging scenario of imbalance and overlapping. Pattern Analysis and Applications, 2008, vol. 11, iss. 3-4, pp. 269–280. URL: Link
Japkowicz N., Stephen S. The class imbalance problem: A systematic study. Intelligent Data Analysis, 2002, vol. 6, no. 5, pp. 429–449 URL: Link
Weiss G.M., Provost F. Learning when training data are costly: The effect of class distribution on tree induction. Journal of Artificial Intelligence Research, 2003, vol. 19, pp. 315–354. URL: Link
Lin Y., Lee Y., Wahba G. Support vector machines for classification in nonstandard situations. Machine Learning, 2002, vol. 46, iss. 1-3, pp. 191–202. URL: Link
Wu G., Chang E. KBA: kernel boundary alignment considering imbalanced data distribution. IEEE Transactions on Knowledge and Data Engineering, 2005, vol. 17, iss. 6, pp. 786–795. URL: Link
Chawla N.V., Bowyer K.W., Hall L.O., Kegelmeyer W.P. SMOTE: synthetic minority over-sampling technique. Journal of Artificial Intelligence Research, 2002, vol. 16, pp. 321–357. URL: Link
He H., Bai Y., Garcia E.A., Li S. ADASYN: Adaptive synthetic sampling approach for imbalanced learning. IEEE International Joint Conference on Neural Networks (IEEE World Congress on Computational Intelligence), 2008, pp. 1322–1328. URL: Link
Han H., Wang W.-Y., Mao B.-H. Borderline-SMOTE: a new over-sampling method in imbalanced data sets learning. In: Huang D.S., Zhang X.P., Huang G.B. (eds) Advances in Intelligent Computing.ICIC 2005. Lecture Notes in Computer Science, 2005, vol. 3644, pp. 878–887. URL: Link
Nguyen H.M., Cooper E.W., Kamei K. Borderline over-sampling for imbalanced data classification. International Journal of Knowledge Engineering and Soft Data Paradigms, 2011, vol. 3, iss. 1, pp. 4–21. URL: Link
Last F., Douzas G., Bacao F. Oversampling for Imbalanced Learning Based on k-Means and SMOTE. URL: Link
Mani I., Zhang I. kNN approach to unbalanced data distributions: a case study involving information extraction. Proceedings of Workshop on Learning from Imbalanced Datasets, 2003. URL: Link
Tomek I. Two modifications of CNN. IEEE Transactions on Systems, Man, and Cybernetics, 1976, vol. SMC-6, iss. 11, pp. 769–772. URL: Link
Wilson D. Asymptotic Properties of Nearest Neighbor Rules Using Edited Data. IEEE Transactions on Systems, Man, and Cybernetics, 1972, vol. SMC-2, iss. 3, pp. 408–421. URL: Link
Hart P. The condensed nearest neighbor rule. IEEE Transactions on Information Theory, 1968, vol. 14, iss. 3, pp. 515–516. URL: Link
Kubat M., Matwin S. Addressing the curse of imbalanced training sets: one-sided selection. Proceedings of the Fourteenth International Conference on Machine Learning, 1997, vol. 97, pp. 179–186.
Smith M.R., Martinez T., Giraud-Carrier C. An instance level analysis of data complexity. Machine Learning, 2014, vol. 95, iss. 2, pp. 225–256. URL: Link
Domingos P. MetaCost: a general method for making classifiers cost-sensitive. Proceedings of the Fifth ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, 1999, pp. 155–164. URL: Link
Ling C.X., Sheng V.S., Yang Q. Test strategies for cost-sensitive decision trees. IEEE Transactions on Knowledge and Data Engineering, 2006, vol. 18, iss. 8, pp. 1055–1067. URL: Link
Schapire R.E. The strength of weak learnability. Machine Learning, 1990, vol. 5, iss. 2, pp. 197–227. URL: Link
Freund Y., Schapire R.E. A decision-theoretic generalization of on-line learning and an application to boosting. Journal of Computer and System Sciences, 1997, vol. 55, iss. 1, pp. 119–139. URL: Link
Schapire R.E., Singer Y. Improved boosting algorithms using confidence-rated predictions. Machine Learning, 1999, vol. 37, iss. 3, pp. 297–336. URL: Link
Breiman L. Bagging predictors. Machine Learning, 1996, vol. 24, iss. 2, pp. 123–140. URL: Link
Aslam J.A., Popa R.A., Rivest R.L. On Estimating the Size and Confidence of a Statistical Audit. Proceedings of the USENIX Workshop on Accurate Electronic Voting Technology, 2007.
Oza N.C., Tumer K. Classifier ensembles: Select real-world applications. Information Fusion, 2008, vol. 9, iss. 1, pp. 4–20. URL: Link
Chawla N.V., Lazarevic A., Hall L.O., Bowyer K.W. SMOTEBoost: Improving prediction of the minority class in boosting. In: Lavrač N., Gamberger D., Todorovski L., Blockeel H. (eds) Knowledge Discovery in Databases: PKDD 2003. PKDD 2003. Lecture Notes in Computer Science, 2003, vol. 2838, Berlin, Springer, pp. 107–119. URL: Link
Seiffert C., Khoshgoftaar T.M., Van Hulse J., Napolitano A. Rusboost: A hybrid approach to alleviating class imbalance. IEEE Transactions on Systems, Man, and Cybernetics – Part A: Systems and Humans, 2010, vol. 40, iss. 1, pp. 185–197. URL: Link
Hu S., Liang Y., Ma L., He Y. MSMOTE: Improving classification performance when training data is imbalanced. Second International Workshop on Computer Science and Engineering, 2009, vol. 2, pp. 13–17. URL: Link
Wang S., Yao X. Diversity analysis on imbalanced data sets by using ensemble models. IEEE Symposium on Computational Intelligence and Data Mining, 2009, pp. 324–331. URL: Link
Tao D., Tang X., Li X., Wu X. Asymmetric bagging and random subspace for support vector machines-based relevance feedback in image retrieval. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2006, vol. 28, iss. 7, pp. 1088–1099. URL: Link
Chang E., Li B., Wu G., Goh K. Statistical learning for effective visual information retrieval. Proceedings 2003 International Conference on Image Processing, 2003, pp. 609–612. URL: Link
Hido S., Kashima H., Takahashi Y. Roughly balanced bagging for imbalanced data. Statistical Analysis and Data Mining, 2009, vol. 2, iss. 5-6, pp. 412–426. URL: Link
Chan P.K., Stolfo S.J. Toward scalable learning with non-uniform class and cost distributions: A case study in credit card fraud detection. Proceedings of the Fourth International Conference on Knowledge Discovery and Data Mining, 1998, pp. 164–168. URL: Link
Yan R., Liu Y., Jin R., Hauptmann A. On predicting rare classes with SVM ensembles in scene classification. 2003 IEEE International Conference on Acoustics, Speech, and Signal Processing, 2003. Proceedings. (ICASSP '03), 2003, vol. 3, pp. 21–24. URL: Link
Liu X.-Y., Wu J., Zhou Z.-H. Exploratory undersampling for class-imbalance learning. IEEE Transactions on Systems, Man, and Cybernetics, Part B (Cybernetics), 2009, vol. 39, iss. 2, pp. 539–550. URL: Link
Fan W., S. Stolfo J., Zhang J., Chan P.K. Adacost: Misclassification cost-sensitive boosting. Proceedings of the Sixteenth International Conference on Machine Learning, 1999, pp. 97–105.
Ting K.M. A comparative study of cost-sensitive boosting algorithms. Proceedings of the Seventeenth International Conference on Machine Learning, 2000, pp. 983–990.
Sun Y., Kamel M.S., Wong A.K., Wang Y. Cost-sensitive boosting for classification of imbalanced data. Pattern Recognition, 2007, vol. 40, iss. 12, pp. 3358–3378. URL: Link
Joshi M.V., Kumar V., Agarwal R.C. Evaluating boosting algorithms to classify rare classes: Comparison and improvements. Proceedings 2001 IEEE International Conference on Data Mining, 2001, pp. 257–264. URL: Link
Estabrooks A., Jo T., Japkowicz N. A multiple resampling method for learning from imbalanced data sets. Computational Intelligence, 2004, vol. 20, iss. 1, pp. 18–36. URL: Link
Stefanowski J., Wilk S. Selective pre-processing of imbalanced data for improving classification performance. In: Song I.Y., Eder J., Nguyen T.M. (eds) Data Warehousing and Knowledge Discovery. DaWaK 2008. Lecture Notes in Computer Science, 2008, vol. 5182, Berlin, Springer, pp. 283–292. URL: Link
Batista G.E.A.P.A., Prati R.C., Monard M.C. A study of the behavior of several methods for balancing machine learning training data. SIGKDD Explorations Newsletter, 2004, vol. 6, iss. 1, pp. 20–29.
Fernandez A., Garcıa S., del Jesus M.J., Herrera F. A study of the behaviour of linguistic fuzzy rule based classification systems in the framework of imbalanced data-sets. Fuzzy Sets and Systems, 2008, vol. 159, iss. 18, pp. 2378–2398. URL: Link
Galar M., Fernandez A., Barrenechea E. et al. A Review on Ensembles for the Class Imbalance Problem: Bagging-, Boosting-, and Hybrid-Based Approaches. IEEE Transactions on Systems, Man, and Cybernetics, Part C (Applications and Reviews), 2012, vol. 42, iss. 4, pp. 463–484. URL: Link