The Korean Journal of Applied Statistics (응용통계연구)

The Korean Statistical Society (한국통계학회)
권호 표지
DOI QR코드

DOI QR Code

KCYP data analysis using Bayesian multivariate linear model

베이지안 다변량 선형 모형을 이용한 청소년 패널 데이터 분석

  • Insun, Lee (Department of Statistics, Sungkyunkwan University) ;
  • Keunbaik, Lee (Department of Statistics, Sungkyunkwan University)
  • Received : 2022.07.05
  • Accepted : 2022.09.05
  • Published : 2022.12.31
Download PDF
⟨ Previous Next ⟩

Abstract

Although longitudinal studies mainly produce multivariate longitudinal data, most of existing statistical models analyze univariate longitudinal data and there is a limitation to explain complex correlations properly. Therefore, this paper describes various methods of modeling the covariance matrix to explain the complex correlations. Among them, modified Cholesky decomposition, modified Cholesky block decomposition, and hypersphere decomposition are reviewed. In this paper, we review these methods and analyze Korean children and youth panel (KCYP) data are analyzed using the Bayesian method. The KCYP data are multivariate longitudinal data that have response variables: School adaptation, academic achievement, and dependence on mobile phones. Assuming that the correlation structure and the innovation standard deviation structure are different, several models are compared. For the most suitable model, all explanatory variables are significant for school adaptation, and academic achievement and only household income appears as insignificant variables when cell phone dependence is a response variable.

다변량 경시적 자료 분석은 반복 측정된 자료에 존재하는 상관관계를 올바르게 추정하면서 자료를 분석해야 한다. 경시적 연구에서는 다변량 경시적 자료가 주로 생성되지만, 기존 통계적 모형은 대부분 단변량으로 분석되어 다변량 경시적 자료에 존재하는 복잡한 상관관계를 제대로 설명하지 못하게 된다. 따라서 본 논문에서는 복잡한 상관관계를 설명하기 위해 공분산 행렬을 모형화하는 다양한 방법에 대해 고찰한다. 그 중 수정된 콜레스키 분해, 수정된 콜레스키 블록분해와 초구분해를 살펴본다. 그리고 일반화 자기회귀모수 행렬이 가지는 희박성 문제를 해결하기 위해 베이지안 방법을 이용하여 청소년 패널 데이터를 분석한다. 청소년 패널 데이터는 다변량 경시적 자료이며, 반응 변수로는 학교 적응도, 학업 성취도, 휴대전화 의존도를 고려한다. 자기 상관 구조와 혁신 표준 편차 구조를 달리 가정하여 여러 모형을 비교한다. 가장 적합한 모형에 대해 학교 적응도와 학업 성취도에 대해 모든 설명 변수가 유의미하며, 휴대전화 의존도가 반응 변수일 때 사교육 시간을 제외한 모든 설명 변수가 유의미한 것으로 나타난다.

Keywords

Acknowledgement

이 논문은 정부의 재원으로 한국연구재단의 지원을 받아 수행된 기초연구사업임 (NRF-2022R1A2C1002752).

References

  1. Kim C and Zimmerman DL (2012). Unconstrained models for the covariance structure of multivariate longitudinal data, Journal of Multivariate Analysis, 107, 104-118. https://doi.org/10.1016/j.jmva.2012.01.004
  2. Chang SJ, Song SW, and Cho MA (2011). The effects of mobile phone dependency, perceived parenting attitude, attachment to peer on school life adjustment of middle school students, Korean Journal of Youth Studies, 18, 431-451.
  3. Chang SJ, Song SW, and Cho MN (2018). The effects of mobile phone dependency, perceived parenting attitude, attachment to peer on school life adjustment of mobile school students, Korean Journal of Youth Studies, 18, 432-451.
  4. Huang JZ, Liu N, Pourahmadi M, and Liu L (2006). Covariance matrix selection and estimation via penalised normal likelihood, Biometrika, 93, 85-98. https://doi.org/10.1093/biomet/93.1.85
  5. Jang SH and Cho KD (2010). Effects of depression scale, interaction anxiety and school adjustment on cellular phone addiction of teenagers, The Journal of the Korea Contents Association, 10, 285-293. https://doi.org/10.5392/JKCA.2010.10.11.285
  6. Kim JA (2016). A longitudinal relationship between adolescents' school adaptation and academic achievement according to parenting attitudes, Korea Journal of Counseling, 17, 303-326. https://doi.org/10.15703/KJC.17.2.201604.303
  7. Kim SY and Hong SH (2014). Estimating adolescent's changes in mobile phone dependence: Testing for the effects of ecological factors on individual differences in the changes, Studies on Korean Youth, 25, 101-123. https://doi.org/10.14816/sky.2014.25.3.101
  8. Kim Y and Lee K (2022). Comparison study for Bayesian multivariate linear model, Journal of the Korean Data & Information Science Society, 33, 249-268. https://doi.org/10.7465/jkdi.2022.33.2.249
  9. Lee K and Cho H, Kwak MS, and Jang EJ (2020). Estimation of covariance matrix of multivariate longitudinal data using modified Cholesky and hypersphere decompositions, Biometrics, 76, 75-86. https://doi.org/10.1111/biom.13113
  10. Lee KJ, Chen RB, Kwak MS, and Lee K (2021). Determination of correlations in multivariate longitudinal data with modified Cholesky and hypersphere decomposition using Bayesian variable selection approach, Statistics in Medicine, 40, 978-997. https://doi.org/10.1002/sim.8815
  11. Pourahmadi M (1999). Joint mean-covariance models with applications to longitudinal data: Unconstrained parameterisation, Biometrika, 86, 677-690. https://doi.org/10.1093/biomet/86.3.677
  12. Pourahmadi M (2000). Maximum likelihood estimation of generalized linear models for multivariate normal covariance matrix, Biometrika, 87, 425-435. https://doi.org/10.1093/biomet/87.2.425
  13. Yoo ES (1996). A study on the change in behavior of children with poor learning through group counseling, Graduate School of Education at Korea University, Master's Thesis, Seoul.
  14. Verbeke G, Fieuws S, Molenberghs G, and Davidian M (2014). The analysis of multivariate longitudinal data: A review, Statistical Methods in Medical Research, 23, 42-59. https://doi.org/10.1177/0962280212445834