The Korean Journal of Applied Statistics (응용통계연구)

The Korean Statistical Society (한국통계학회)
권호 표지
DOI QR코드

DOI QR Code

Discriminant analysis for unbalanced data using HDBSCAN

불균형자료를 위한 판별분석에서 HDBSCAN의 활용

  • Lee, Bo-Hui (Department of Advertising and Public Relations, Silla University) ;
  • Kim, Tae-Heon (Department of Statistics, Pusan National University) ;
  • Choi, Yong-Seok (Department of Statistics, Pusan National University)
  • Received : 2021.04.05
  • Accepted : 2021.04.14
  • Published : 2021.08.31
Download PDF
⟨ Previous Next ⟩

Abstract

Data with a large difference in the number of objects between clusters are called unbalanced data. In discriminant analysis of unbalanced data, it is more important to classify objects in minority categories than to classify objects in majority categories well. However, objects in minority categories are often misclassified into majority categories. In this study, we propose a method that combined hierarchical DBSCAN (HDBSCAN) and SMOTE to solve this problem. Using HDBSCAN, it removes noise in minority categories and majority categories. Then it applies SMOTE to create new data. Area under the roc curve (AUC) and F1 scores were used to compare performance with existing methods. As a result, in most cases, the method combining HDBSCAN and synthetic minority oversampling technique (SMOTE) showed a high performance index, and it was found to be an excellent method for classifying unbalanced data.

군집간의 개체 수의 차이가 큰 자료들을 불균형자료라고 한다. 불균형자료의 판별분석에서 다수 범주의 개체를 잘 분류하는 것 보다 소수 범주의 개체를 잘 분류하는 것이 더 중요하다. 그러나 개체 수가 상대적으로 작은 소수 범주의 개체를 개체 수가 상대적으로 많은 다수 범주의 개체로 오분류하는 경우가 많다. 본 연구에서는 이를 해결하기 위해 HDBSCAN과 SMOTE를 결합한 방법을 제안한다. HDBSCAN을 이용하여 소수 범주의 노이즈와 다수 범주의 노이즈를 제거하고 SMOTE를 적용하여 새로운 자료를 만들어낸다. 기존의 방법들과 성능을 비교하기 위하여 AUC와 F1 점수를 이용하였고 그 결과 대부분의 경우에 HDBSCAN과 SMOTE를 결합한 방법이 높은 성능 지표를 보였고, 불균형자료를 분류하는데 있어 뛰어난 방법으로 나타났다.

Keywords

Acknowledgement

본 논문은 교육부 및 한국연구재단의 4단계 두뇌한국(BK)21 사업으로 지원된 연구임.

References

  1. Chawla NV, Hall LO, Bowyer KW, and Kegelmeyer WP (2002). Smote: Synthetic minority oversampling technique, Journal of Artificial Intelligence Research, 16, 321-357. https://doi.org/10.1613/jair.953
  2. Choi YS (2018). Multivariate Data Analysis with R, Kyungmoon, Seoul.
  3. Han H, Wang W, and Mao B (2005). Borderline smote: Anew over sampling method in imbalanced data sets learning. In Proceedings of International Conference on Intelligent Computing, 878-887.
  4. He H, Bai Y, Garcia EA, and Li S (2008). ADASYN: Adaptive synthetic sampling approach for imbalanced learning. In Proceedings of International Joint Conference on Neural Networks, 1322-1328.
  5. Ijaz M, Alfian G, Syafrudin M, and Rhee J (2018). Hybrid prediction model for type 2 diabetes and hypertension using DBSCAN-based outlier detection, synthetic minority over sampling technique (SMOTE), and Random Forest, Applied Sciences, 8, 1325. https://doi.org/10.3390/app8081325
  6. McInnes L and Healy J (2017). Accelerated hierarchical density based clustering, IEEE International Conference on Data Mining Workshops (ICDMW).