Bayesian negative binomial logit hurdle and zero-inflated model for characterizing smoking intensity

Abstract Smoking invariably has environmental, social, economic and health consequences in Ethiopia. Reducing quitting cigarette smoking improves individual increases available household funds for education, food better productivity. Therefore, this study aimed to apply the Bayesian negative binomial logit hurdle zero-inflated model determine associated factors of number smokers per day using intensity data 2016 Ethiopia Demographic Health Survey. The survey was a community-based cross-sectional conducted from January 18 June 27, 2016. used two stage stratified sampling design. analysis Negative Binomial Logit Hurdle Zero-inflated models which incorporate both overdispersion excess zeros carry out estimation Markov Chain Monte Carlo techniques. About 94.2% them never cigarettes smoked were found have overdispersion. after considering zero counts enduring overdispersion, according AIC Vuong tests, best fit data. finding technique is more robust precisely due that it popular method. Furthermore; Zero-inflation Zero variable: age, residence, education level, internet use, wealth index, marital status, chewed chat, occupation, media most statistically significant determinate on intensity.