Search for


TEXT SIZE
search for




 

동적 베이지안 네트워크 모형을 이용한 한국 주식시장의 동시적 및 시간적 연관관계 연구
Discovering contemporaneous and temporal dependence structure of Korean stock market using a dynamic Gaussian Bayesian network
Korean J Appl Stat 2024;37(6):721-732
Published online December 31, 2024
© 2024 The Korean Statistical Society.

박소연a, 오로지1,b, 오만숙2,a
So-Yeon Parka, Rosy Oh1,b, Man-Suk Oh2,a

a이화여자대학교 통계학과; b육군사관학교 수학과

aDepartment of Statistics, Ewha Womans University;
bDepartment of Mathematics, Korea Military Academy
1Corresponding author: Department of Mathematics, Korea Military Academy, PO Box 77-2, Nowon-gu, Seoul 01805, Korea. E-mail: rosy.oh5@gmail.com
2Corresponding author: Department of Statistics, Ewha Womans University, 52 Ewhayeodae-gil, Seodaemun-gu, Seoul 03760,Korea. E-mail: msoh@ewha.ac.kr
This work was supported by the National Research Foundation of Korea (NRF) grant funded by the Korean government (MSIT) (NRF-2022R1A2C1091256). This paper was prepared by extracting part of So-Yeon Park’s thesis.
Received February 29, 2024; Revised April 9, 2024; Accepted April 15, 2024.
Abstract
이 연구는 동적 가우시안 베이지안 네트워크(dynamic Gaussian Bayesian network; DGBN) 모형을 이용하여 한국 주식시장에 영향을 미치는 요인들 간의 동시적, 시간적 연관 구조를 탐색한다. DGBN은 기존의 베이지안 네트워크를 확장하여 시계열 자료에서 동일시점 뿐만 아니라 시간에 따른 연관성을 그래프로 보여주는 머신러닝 기법의 일종이다. 2015년부터 2022년까지의 기간동안 관측된 한국 주식시장 관련 국내외 요인들의 관측값을 사용하여 학습된 DGBN을 통하여 동일한 주식 개장일에 어떻게 국내외 요인들이 서로 상호작용 하는지 알 수 있었다. 또한, 한국 주식시장에서 주식 개장일 기준 하루 전 요인 관측값이 현재 일자 자신을 포함한 전체 요인에게 어떻게 영향을 주는지 보여주었다. 이같은 결과는 국내외 경제적 요인들이 한국주식시장에 미치는 영향을 이해하는데 중요한 인사이트를 제공하며, 이는 개인 투자자의 투자 결정과 향후 금융 시장 및 경제 정책 수립에 유용한 정보를 제공할 것으로 예상된다.
This study investigates the dependence structure among the factors of Korean stock market using a dynamic Gaussian Bayesian network (DGBN). DGBN is an extension of static Gaussian Bayesian network to account for not only contemporaneous dependence but also temporal dependence that may present in the data obverved over time. The DGBN structure learned from data observed from 2015 to 2022 showed how domestic and global factors of Korean stock market interact each other in one calendar day. It also showed how the factors in previous day a ect the present day factors. These results would provide important insights to the e ects of domestic and global factors on Korean stock market.
주요어 : 동적 베이지안 네트워크, 그래프 모형, 주식시장, KOSPI, 시간적 연관성
Keywords : dynamic Bayesian network, graphical model, stock market, KOSPI, temporal dependence
References
  1. Çambas¸ı H, Kuru Ö, Amasyalı MF, and Tahar S (2019). Comparison of Dynamic Bayesian Network Tools, In 2019 Innovations in Intelligent Systems and Applications Conference (ASYU) 2019 Oct 31 (pp. 1-6). IEEE.
    CrossRef
  2. Chang J, Bai Y, Xue J et al. (2023). Dynamic Bayesian networks with application in environmental modeling and management: A review, Environmental Modelling & Software, 2023 Sep 26:105835.
    CrossRef
  3. Cho DH, Moon SH, and Kim YH (2020). A daily stock index predictor using feature selection based on a genetic wrapper, Proceedings of the 2020 Genetic and Evolutionary Computation Conference Companion, Jul 8, 31-32.
    CrossRef
  4. Cho DH, Moon SH, and Kim YH (2021). Genetic feature selection applied to KOSPI and cryptocurrency price prediction, Mathematics, 9, 2574.
    CrossRef
  5. Chung H and Shin KS (2020). Genetic algorithm-optimized multi-channel convolutional neural network for stock market prediction, Neural Computing and Applications, 32, 7897-914.
    CrossRef
  6. Chung HK and SongW(2023). Impacts of COVID-19 on Various Financial Assets in Korea: A Bayesian Network Approach, Available at SSRN 4333917
    CrossRef
  7. Dagum P, Galper A, and Horvitz E (1992). Dynamic network models for forecasting, Proceedings of the Eighth Conference on Uncertainty in Artificial Intelligence, AUAI Press, 41-48.
    CrossRef
  8. Friedman N, Goldszmidt M, and Wyner A (2013). Data analysis with Bayesian networks: A bootstrap approach, arXiv preprint arXiv:1301.6695.
  9. Han JJ and Kim HJ (2023). Prediction of investor-specific trading trends in South Korean stock markets using a BiLSTM prediction model based on sentiment analysis of financial news articles, Journal of Behavioral Finance, 24, 398-410.
    CrossRef
  10. Jo M, Oh R, and Oh M (2022). Prediction of PM10 concentration in Seoul, Korea using Bayesian Network, Communications for Statistical Applications and Methods, 30, 517-530.
    CrossRef
  11. Kyrimi E, McLachlan S, Dube K, Neves MR, Fahmi A, and Fenton N (2021). A comprehensive scoping review of Bayesian networks in healthcare: Past, present and future, Artificial Intelligence in Medicine, 117, 102108.
    Pubmed CrossRef
  12. Korea Securities Depository (2022).
    Available from: https://www.ksd.or.kr/ko/information- disclosure/resourcecenter/periodicals/deposit-and-settlement-form, , No.125
  13. Künsch HR (1989). The jackknife and the bootstrap for general stationary observations, The Annals of Statistics, 17, 1217-1241.
    CrossRef
  14. Lee D and Kwon K (2023). Dynamic Bayesian network model for comprehensive risk analysis of fatigue-critical structural details, Reliability Engineering & System Safety, Jan 1;229:108834.
    CrossRef
  15. Liu Y, Feng H, and Guo K (2021). The dynamic relationship between macroeconomy and stock market in China: Evidence from Bayesian network, Complexity, 2021 Dec 16;2021:1-2.
    CrossRef
  16. Malagrino LS, Roman NT, and Monteiro AM (2018). Forecasting stock market index daily direction: A Bayesian Network approach, Expert Systems with Applications, 105, 11-22.
    CrossRef
  17. Oluseun Olayungbo D, Al-Faryan MA, and Zhuparova A (2023). Network granger causality linkages in Nigeria and developed stock markets: Bayesian graphical analysis, Journal of African Business, 17, 1-25.
    CrossRef
  18. Park J and Kim D (2021). How do Bitcoin and gold affect KOSPI? - focusing on Bayesian network model analysis, Journal of New Industry and Business, 39, 47-64.
  19. Pearl J (1988). Probabilistic reasoning in intelligent systems: networks of plausible inference, Morgan kaufmann.
    CrossRef
  20. Pearl J (2009). Causality, Cambridge University Press, Cambridge.
  21. Pourret O, Na P, and Marcot B (2008). Bayesian networks: A Practical Guide to Applications, John Wiley & Sons, West Sussex.
    Pubmed CrossRef
  22. Quesada D (2022). dbnR: Dynamic Bayesian Network Learning and Inference. r package version 0.7.8. https://cran.r-project.org/web/packages/dbnR/index.html.
  23. Scutari M and Denis JB (2021). Bayesian Networks: With Examples in R, CRC press, New York.
    CrossRef
  24. Scutari M, Kerob D, and Salah S (2023). Inferring Skin-Brain-Skin Connections from Infodemiology Data using Dynamic Bayesian Networks, medRxiv preprint doi: https://doi.org/10.1101/2023.05.15.23290003.
    CrossRef
  25. Tsamardinos I, Brown LE, and Aliferis CF (2006). The max-min hill-climbing Bayesian network structure learning algorithm, Machine Learning, 65, 31-78.
    CrossRef
  26. Yoon I (2016). Impact of macroeconomic indicators on Korea’s stock markets, Regional Industry Review, 39, 37-53.
December 2024, 37 (6)
Full Text(PDF) Free