The Korean Journal of Applied Statistics (응용통계연구)

The Korean Statistical Society (한국통계학회)
권호 표지
DOI QR코드

DOI QR Code

The performance of Bayesian network classifiers for predicting discrete data

이산형 자료 예측을 위한 베이지안 네트워크 분류분석기의 성능 비교

  • 박현재 (중앙대학교 응용통계학과) ;
  • 황범석 (중앙대학교 응용통계학과)
  • Received : 2020.02.04
  • Accepted : 2020.02.29
  • Published : 2020.06.30
Download PDF
⟨ Previous Next ⟩

Abstract

Bayesian networks, also known as directed acyclic graphs (DAG), are used in many areas of medicine, meteorology, and genetics because relationships between variables can be modeled with graphs and probabilities. In particular, Bayesian network classifiers, which are used to predict discrete data, have recently become a new method of data mining. Bayesian networks can be grouped into different models that depend on structured learning methods. In this study, Bayesian network models are learned with various properties of structure learning. The models are compared to the simplest method, the naïve Bayes model. Classification results are compared by applying learned models to various real data. This study also compares the relationships between variables in the data through graphs that appear in each model.

방향성 비순환 그래프(directed acyclic graph; DAG)라고도 하는 베이지안 네트워크(Bayesian network)는 변수 사이의 관계를 확률과 그래프를 통해 모형화할 수 있다는 점에서 최근 의학, 기상학, 유전학 등 여러 분야에서 다양하게 활용되고 있다. 특히 이산형 자료의 예측에 사용되는 베이지안 네트워크 분류분석기(Bayesian network classifier)가 최근 새로운 데이터 마이닝 기법으로 주목받고 있다. 베이지안 네트워크는 그 구조와 학습 방법에 따라 여러 가지 다양한 모형으로 분류할 수 있다. 본 논문에서는 서로 다른 성질을 가진 이산형 자료를 바탕으로 구조 학습 방법에 차이를 두어 베이지안 네트워크 모형을 학습시킨 후, 가장 간단한 방법인 나이브 베이즈 (naïve Bayes) 모형과 비교해 본다. 학습된 모형들을 여러 가지 실제 데이터에 적용하여 그 예측 정확도를 비교함으로써 최적의 분류 분석 결과를 얻을 수 있는지 살펴본다. 또한 각각의 모형에서 나타나는 그래프를 통해 데이터의 변수 사이의 관계를 비교한다.

Keywords

References

  1. Airoldi, E. M. (2007). Getting started in probabilistic graphical models, PLoS Computational Biology, 3, 2421-2425. https://doi.org/10.1371/journal.pcbi.0030252
  2. Bielza, C. and Larranaga, P. (2014). Discrete Bayesian network classifiers: a survey, ACM Computing Surveys, 47, 1-43. https://doi.org/10.1145/2576868
  3. Cheng, J. and Greiner, R. (1999). Comparing Bayesian network classifiers. In Proceedings of the 15th Conference on Uncertainty in Artificial Intelligence, Morgan Kaufmann Publishers, 101-108.
  4. Chow, C. and Liu, C. (1968). Approximating discrete probability distributions with dependence trees, IEEE Transactions On Information Theory, 14, 462-467. https://doi.org/10.1109/TIT.1968.1054142
  5. Epskamp, S., Waldorp, L. J., Mottus, R., and Borsboom, D. (2018). The Gaussian graphical model in cross-sectional and time-series data, Multivariate Behavioral Research, 53, 453-480. https://doi.org/10.1080/00273171.2018.1454823
  6. Friedman, N., Geiger, D., and Goldszmid, M. (1997). Bayesian network classifiers, Machine Learning, 29, 131-163. https://doi.org/10.1023/A:1007465528199
  7. Hand, D. J. and Yu, K. (2001). Idiot's Bayes - not so stupid after all?, International Statistical Review, 69, 385-398. https://doi.org/10.1111/j.1751-5823.2001.tb00465.x
  8. Hansen, K. D., Gentry, J., Long, L., Gentleman, R., Falcon, S., Hahne, F., and Sarkar, D. (2017). Rgraphviz: Provides plotting capabilities for R graph objects. URL https://doi.org/10.18129/B9.bioc.Rgraphviz. R package version 2.20.0.
  9. Horny, M. (2014). Bayesian networks (Technical Report), Boston University School of Public Health.
  10. Keogh, E. J. and Pazzani, M. J. (2002). Learning the structure of augmented Bayesian classifiers, International Journal on Artificial Intelligence Tools, 11, 587-601. https://doi.org/10.1142/S0218213002001052
  11. Leung, D., Drton, M., and Hara, H. (2016). Identifiability of directed Gaussian graphical models with one latent source, Electronic Journal of Statistics, 10, 394-422. https://doi.org/10.1214/16-EJS1111
  12. Margaritis, D. (2003). Learning Bayesian network model structure from data. Ph.D thesis, Carnegie Mellon University, School of Computer Science.
  13. Mihaljevic, B., Bielza, C., and Larranaga, P. (2018). bnclassify: Learning Discrete Bayesian Network Classifiers from Data. URL https://CRAN.R-project.org/package=bnclassify. R package version 0.4.0.
  14. Park, G. (2019). Discovering a fine dust pathway via directed acyclic graphical models, Journal of the Korean Data & Information Science Society, 30, 67-76. https://doi.org/10.7465/jkdi.2019.30.1.67
  15. Rish, I. (2001). An empirical study of the naive Bayes classifier, IJCAI 2001 workshop on empirical methods in artificial intelligence, 3, 41-46.
  16. Robins, G., Pattison, P., Kalish, Y., and Lusher, D. (2007). An introduction to exponential random graph ($p^{\ast}$) models for social networks, Social Networks, 29, 173-191. https://doi.org/10.1016/j.socnet.2006.08.002
  17. Sahami, M. (1996). Learning limited dependence Bayesian classifiers. In Proceedings of the 2nd International Conference on Knowledge Discovery and Data Mining, 335-338.
  18. Webb, G. I., Boughton, J., and Wang, Z. (2005). Not so naive Bayes: Aggregating one-dependence estimators, Machine Learning, 58, 5-24. https://doi.org/10.1007/s10994-005-4258-6
  19. Zheng, F. and Webb, G.I. (2006). Efficient lazy elimination for averaged one-dependence estimators. In Proceedings of the 23rd International Conference on Machine Learning, 148, 1113-1120.