Abstract
It is aimed to investigate the zero inflated regression models and hurdle regression models for describing demodex counts by various variables. It is also aimed to determine the model that will give the best results in describing demodex counts by different variables such as age, sex and profession with the information obtained from patients who come for porter examination. Data were obtained from 156 patients who come to Malatya Public Health Laboratory for porter examination between the dates of June 2007 - June 2009. Demodex count was determined as response variable and age, sex and profession of patients were determined as explanatory variables. R.2.ll.1 software program was used for statistical analysis. Since 68 percent of demodex counts, which was considered as response variable, were zero, zero-inflated regression and hurdle models were used. As a result of applications, AIC values, which were used to determine the superiority of models relative to each other, were obtained as 731.18, 731.49, 531.73 and 531.11 for ZIP, ZAP, ZINB and ZANB models respectively. The model that has smallest AIC value was ZANB. In zero inflated and hurdle regression models, it is tested whether zero values have a significant impact. If the impact of zero values is not significant, the results of zero inflated and hurdle regression models and the results of Poisson regression and negative binomial regression models will be same. In this study, the impact of zero values in the existing data set war determined to be significant.