불균형 데이터 분류를 위한 딥러닝 기반 오버샘플링 기법 = A Deep Learning Based Over-Sampling Scheme for Imbalanced Data Classification

국문 초록 (Abstract)

분류 문제는 주어진 입력 데이터에 대해 해당 데이터의 클래스를 예측하는 문제로, 자주 쓰이는 방법 중의 하나는 주어진 데이터셋을 사용하여 기계학습 알고리즘을 학습시키는 것이다. 이런 경우 분류하고자 하는 클래스에 따른 데이터의 분포가 균일한 데이터셋이 이상적이지만, 불균형한 분포를 가지고 경우 제대로 분류하지 못하는 문제가 발생한다. 이러한 문제를 해결하기 위해 본 논문에서는 Conditional Generative Adversarial Networks(CGAN)을 활용하여 데이터 수의 균형을 맞추는 오버샘플링 기법을 제안한다. CGAN은 Generative Adversarial Networks(GAN)에서 파생된 생성 모델로, 데이터의 특징을 학습하여 실제 데이터와 유사한 데이터를 생성할 수 있다. 따라서 CGAN이 데이터 수가 적은 클래스의 데이터를 학습하고 생성함으로써 불균형한 클래스 비율을 맞추어 줄 수 있으며, 그에 따라 분류 성능을 높일 수 있다. 실제 수집된 데이터를 이용한 실험을 통해 CGAN을 활용한 오버샘플링 기법이 효과가 있음을 보이고 기존 오버샘플링 기법들과 비교하여 기존 기법들보다 우수함을 입증하였다.

다국어 초록 (Multilingual Abstract)

Classification problem is to predict the class to which an input data belongs. One of the most popular methods to do this is training a machine learning algorithm using the given dataset. In this case, the dataset should have a well-balanced class distribution for the best performance. However, when the dataset has an imbalanced class distribution, its classification performance could be very poor. To overcome this problem, we propose an over-sampling scheme that balances the number of data by using Conditional Generative Adversarial Networks (CGAN). CGAN is a generative model developed from Generative Adversarial Networks (GAN), which can learn data characteristics and generate data that is similar to real data. Therefore, CGAN can generate data of a class which has a small number of data so that the problem induced by imbalanced class distribution can be mitigated, and classification performance can be improved. Experiments using actual collected data show that the over-sampling technique using CGAN is effective and that it is superior to existing over-sampling techniques.

참고문헌 (Reference)

1 I. Tomek, "Two Modifications of CNN" 6 : 769-772, 1976

2 C. Cortes, "Support-vector networks" 20 (20): 273-297, 1995

3 N. V. Chawla, "SMOTE : synthetic minority oversampling technique" 16 : 321-357, 2002

4 V. Nair, "Rectified linear units improve restricted boltzmann machines" 807-814, 2010

5 G. W. Corder, "Nonparametric Statistics for Non-Statisticians: A Step-by-Step Approach" John Wiley & Sons 2014

6 S. S. Haykin, "Neural networks and learning machines, Vol. 3" Pearson 2009

7 "MineThatData"

8 G. Haixiang, "Learning from class-imbalanced data : Review of methods and applications" 73 : 220-239, 2017

9 I. Jolliffe, "International Encyclopedia of Statistical Science" Springer 1094-1096, 2011

10 I. J. Goodfellow, "Generative adversarial nets" 2672-2680, 2014

11 X. Y. Liu, "Exploratory undersampling for class-imbalance learning" 39 (39): 539-550, 2009

12 J. Van Hulse, "Experimental perspectives on learning from imbalanced data" 935-942, 2007

13 C. X. Ling, "Data mining for direct marketing:Problems and solutions" 73-79, 1998

14 S. J. Salvatore, "Cost-based modeling for fraud and intrusion detection : Results from the JAM project" 130-144, 2000

15 M. Mirza, "Conditional generative adversarial nets"

16 H. Y. Wang, "Combination Approach of SMOTE and Biased-SVM for Imbalanced Datasets" 228-231, 2008

17 A. Liaw, "Classification and regression by randomForest" 2 (2): 18-22, 2002

18 T. Jo, "Class Imbalances versus Small Disjuncts" 6 : 40-49, 2004

19 H. Han, "Borderline-SMOTE:a new over-sampling method in imbalanced data sets learning" 878-887, 2005

20 H. M. Nguyen, "Borderline Over-Sampling for Imbalanced Data Classification" 3 (3): 4-21, 2011

21 Y. Liu, "Boosting Prediction Accuracy on Imbalanced Datasets with SVM Ensembles" 107-118, 2006

22 C. H. Hoi, "Biased support vector machine for relevance feedback in image retrieval" 2004

23 D. L. Wilson, "Asymptotic Properties of Nearest Neighbor Rules Using Edited Data" 3 : 408-421, 1972

24 A. Dal Pozzolo, "Adaptive machine learning for credit card fraud detection" 2015

25 D. P. Kingma, "Adam: A method for stochastic optimization" 2014

26 H. He, "ADASYN: Adaptive synthetic sampling approach for imbalanced learning" 1322-1328, 2008

27 N. Macia, "A.Preliminary Approach in Synthetic Data Sets Generation based on Class Separability Measure" 1-4, 2008

28 R. O’Brien, "A random forests quantile classifier for class imbalanced data" 90 : 232-249, 2019

29 S. Moro, "A data-driven approach to predict the success of bank telemarketing" 62 : 22-31, 2014

30 G. E. Batista, "A Study of the Behavior of Several Methods for Balancing Machine Learning Training Data" 6 (6): 20-29, 2004