Comparing Bayesian regression and classic zero-inflated negative binomial on size estimation of people who use alcohol
Abstract
Background & Aim: Nowadays, we have some data in different sciences which number of zeros is more than expected, such data are called zero-inflated which can be modeled by regressions for count data. Many researches have been conducted in the field of classical method on count data. Most of Bayesian analysis which is conducted for these data used zero-inflated Poisson regression. Therefore, the main purpose of this research is comparison of Bayesian and classic approaches in regression of zero-inflated negative binomial (NB) on data for determining the size estimation of people who have used alcohol more than once in last year.
Methods & Materials: This research had been in two provinces of Fars and Kerman in 2011, a sample size of each province was formed proportional to people of that province, and totally the calculated sample size was 700. Zero-inflated NB regression was fitted to the data in two Bayesian and classical methods, and then two methods have been compared. Results of Bayesian method were extracted in OpenBUGS software and through related codes in R and results of classical method were extracted in R software too.
Results: After fitting classical method, variables of province, gender, age groups, and education had been effective on identifying number of alcoholics, but in Bayesian method, three variables of gender, age groups, and education have become significant. In this research, it was specified that obtained probability intervals from Bayesian method are much Widder than classical method.
Conclusion: Results of this research indicate that Bayesian method has better function than the classic
Bohning D, Dietz E, Schlattmann P, Mendonca L, Kirchner U. The zero-inflated poisson model and the decayed, missing and filled teeth index in dental epidemiology. J R Stat Soc Ser A Stat Soc 1999; 162(2): 195-209.
Shankar V, Milton J, Mannering F. Modeling accident frequencies as zero-altered probability processes: An empirical inquiry. Accid Anal Prev 1997; 29(6): 829-37.
Zhou XH, Tu W. Confidence intervals for the mean of diagnostic test charge data containing zeros. Biometrics 2000; 56(4): 1118-25. 4. Agarwal DK, Gelfand AE, Citron-Pousty S. Zero-inflated models with application to spatial count data. Environ Ecol Stat 2002; 9(4): 341-55.
Hall DB. Zero-inflated poisson and binomial regression with random effects: A case study. Biometrics 2000; 56(4): 1030-9.
Heilbron D, Gibson D. Shared needle use and health beliefs concerning AIDS: Regression modeling of zero-heavy count data. Proceedings of the 6th International Conference on AIDS that was held; 1990 June 20-24; San Francisco, CA.
Lambert D. Zero-inflated poisson regression, with an application to defects in manufacturing. Technometrics 1992; 34(1): 1-14.
Famoye F. Restricted generalized poisson regression model. Commun Stat Theory Methods 1993; 22(5): 1335-54.
Lawless JF. Negative binomial and mixed Poisson regression. Can J Stat 1987; 15(3): 209-25.
Ozmen I, Demirhan H. A bayesian approach for zero-inflated count regression models by using the reversible jump markov chain monte carlo method and an application. Commun Stat Theory Methods 2010; 39(3): 2109-27.
Zaninotto P, Falaschetti E. Comparison of methods for modelling a count outcome with excess zeros: Application to Activities of Daily Living (ADL-s). J Epidemiol Community Health 2011; 65(3): 205-10. 12. Bolstad WM. Introduction to bayesian statistics. 2nd ed. New York, NY: John Wiley & Sons; 2013.
Neal DJ, Sugarman DE, Hustad JT, Caska CM, Carey KB. It's all fun and games...or is it? Collegiate sporting events and celebratory drinking. J Stud Alcohol 2005; 66(2): 291-4.
Steers ML, Coffman AD, Wickham RE, Bryan JL, Caraway L, Neighbors C. Evaluation of alcohol-related personalized normative feedback with and without an injunctive message. J Stud Alcohol Drugs 2016; 77(2): 337-42.
Bloomfield K, Gmel G, Wilsnack S. Introduction to special issue 'Gender, Culture and Alcohol Problems: A Multi-national Study'. Alcohol Alcohol Suppl 2006; 41(1): i3-i7.
Jang H, Lee S, Kim SW. Bayesian analysis for zero-inflated regression models with the power prior: Applications to road safety countermeasures. Accid Anal Prev 2010; 42(2): 540-7.
Ghosh SK, Mukhopadhyay P, Lu JC. Bayesian analysis of zero-in?ated regressionmodels. J Stat Plan Inference 2017; 136(4): 1360-75.
Liu GF, Han B, Zhao X, Lin Q. A Comparison of frequentist and bayesian model based approaches for missing data analysis: Case study with a schizophrenia clinical trial. Stat Biopharm Res 2016; 8(1): 116-27.
Yu J, Hutson AD, Siddiqui AH, Kedron MA. Group sequential control of overall toxicity incidents in clinical trials-non-Bayesian and Bayesian approaches. Stat Methods Med Res 2016; 25(1): 64-80. 20. Fosu MO, Jackson OA, Twum SB. Bayesian and frequentist comparison: An application to low birth weight babies in Ghana. Br J Appl Sci Technol 2016; 16(2): 1-15.
Nikfarjam A, Shokoohi M, Shahesmaeili A, Haghdoost AA, Baneshi MR, HajiMaghsoudi S, et al. National population size estimation of illicit drug users through the network scale-up method in 2013 in Iran. Int J Drug Policy 2016; 31: 147-52.
Toft N, Innocent GT, Gettinby G, Reid SWJ. Assessing the convergence of markov chain monte carlo methods: An example from evaluation of diagnostic tests in absence of a gold standard. Prev Vet Med 2007; 79(2): 244-56.
Jalali M, Nikfarjam A, Haghdoost AA, Memaryan N, Tarjoman T, Baneshi MR. Social hidden groups size analyzing: Application of count regression models for excess zeros. J Res Health Sci 2013; 13(2): 143-50.
| Files | ||
| Issue | Vol 2 No 4 (2016) | |
| Section | Original Article(s) | |
| Keywords | ||
| Bayesian Markov chain Monte Carlo method Zero-inflated Bayesian analysi | ||
| Rights and permissions | |
|
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License. |
