DOI QR코드

DOI QR Code

Bayesian Analysis of a Zero-inflated Poisson Regression Model: An Application to Korean Oral Hygienic Data

영과잉 포아송 회귀모형에 대한 베이지안 추론: 구강위생 자료에의 적용

  • 임아경 (이화여자대학교 통계학과) ;
  • 오만숙 (이화여자대학교 통계학과)
  • Published : 2006.11.30

Abstract

We consider zero-inflated count data, which is discrete count data but has too many zeroes compared to the Poisson distribution. Zero-inflated data can be found in various areas. Despite its increasing importance in practice, appropriate statistical inference on zero-inflated data is limited. Classical inference based on a large number theory does not fit unless the sample size is very large. And regular Poisson model shows lack of St due to many zeroes. To handle the difficulties, a mixture of distributions are considered for the zero-inflated data. Specifically, a mixture of a point mass at zero and a Poisson distribution is employed for the data. In addition, when there exist meaningful covariates selected to the response variable, loglinear link is used between the mean of the response and the covariates in the Poisson distribution part. We propose a Bayesian inference for the zero-inflated Poisson regression model by using a Markov Chain Monte Carlo method. We applied the proposed method to a Korean oral hygienic data and compared the inference results with other models. We found that the proposed method is superior in that it gives small parameter estimation error and more accurate predictions.

셀 수 있는 이산 자료(discrete count data)에 대한 분석은 여러 분야에서 활용되고 있지만 영(zero)을 과도하게 포함하고 있는 영과잉 자료는 자료의 성격상 포아송 분포를 따르지 못할 때가 있어 분석에 어려움이 따른다. Zero-Inflated Poisson(ZIP)모형은 이런 어려움을 극복하기 위하여 영에 대한 점확률을 가지는 분포와 포아송 분포를 합성하여 과도한 영과 영이 아닌 자료를 설명하는 모형이다. 설명 변수가 존재할 때는 포아송 분포 부분에서 반응변수의 평균과 공변량사이에 로그선형 연결함수를 사용한 Zero-Inflated Poisson Regression(ZIPR)모형이 사용될 수 있다. 본 논문에서는 Markov Chain Monte Carlo 기법을 이용한 ZIPR모형의 베이지안 추론방법을 제안하고, 이를 실제 구강위생 자료에 적용하며 다른 모형들과 비교한다. 그 결과 베이지안 추론 방법을 적용한 영과잉 모형의 추정오차가 다른 모형들의 추정오차보다 작았고, 예측치가 더 정확했다는 점에서 우수함을 알 수 있었다.

Keywords

References

  1. Angers, J. F. and Biswas, A. (2003). Bayesian Analysis of Zero-Inflated generalized Poisson model, Computational Statistics and Data Analysis 42, 37-46 https://doi.org/10.1016/S0167-9473(02)00154-8
  2. Dahiya, R. C. and Gross, A. J. (1973). Estimating the zero class from a truncated poisson sample, Journal of the American Statistical Association 68, 731-733 https://doi.org/10.1080/01621459.1973.10481415
  3. Ghosh S. K. and Mukhopadhyay P. (2005). Bayesian analysis of zero-inflated regression models, Journal of statistical planning and inference In press, 1-16
  4. Gupta, P. L. and Gupta, R. C. and Tripathi R. C. (1996). Analysis of zero-adjusted count data, Computational Statistics and Data Analysis 23, 207-218 https://doi.org/10.1016/S0167-9473(96)00032-1
  5. Lambert, D. (1992). Zero-Inflated Poisson Regression With an application to Defects in Manufacturing, Technometrics 34, 1-14 https://doi.org/10.2307/1269547
  6. Li, C. S., Lu, J. C.,Park, J., Kim, K. M., Brinkley, P. A., Peterson, J., (1999). A multivariate zero-inflated Poisson distribution and its inference, Technometrics 41, 29-38 https://doi.org/10.2307/1270992
  7. Lord, D. and Washington, S. P. and Ivan, J. N.(2005). Poisson, Poisson-gamma and zeroinflated regression models of motor vehicle crashes:balancing statistical fit and theory, Accident Analysis and Prevention 37, 35-46 https://doi.org/10.1016/j.aap.2004.02.004
  8. Rodrigues, R. (2003). Bayesian Analysis of Zero-Inflated Distributions, Communications in Statistics 32, 281-289 https://doi.org/10.1081/STA-120018186
  9. Umbach, D. (1981). On inference for a mixture of Poisson and a degenerate distribution, Communications in Statistics: theory and methods 10, 299-306 https://doi.org/10.1080/03610928108828039
  10. Yip, P. (1988). Inference about the mean of a Poisson distribution in the presence of nuisance parameter, Australian Journal of Statistics 30, 299-306 https://doi.org/10.1111/j.1467-842X.1988.tb00624.x

Cited by

  1. Analysis of Elderly Drivers' Accident Models Considering Operations and Physical Characteristics vol.30, pp.6, 2012, https://doi.org/10.7470/jkst.2012.30.6.037
  2. Bayesian Inference for the Zero In ated Negative Binomial Regression Model vol.24, pp.5, 2011, https://doi.org/10.5351/KJAS.2011.24.5.951