Analysis Of A Class Of Infectious Diseases Based On The Zero Inflated Model

Tuberculosis is an infectious disease that seriously endangers people's health.School is also a place of high density of people,but also prone to tuberculosis.In 1992,the hepatitis B virus was incorporated into the scope of planned immunity in China,which make the school tuberculosis prevention work has made great progress.In order to find out the influencing factors of the case of tuberculosis,the research work of the text is carried out.This paper analyzes the influence factors of tuberculosis transmission data in Dalian school from 2013-2015.Regression models and zero inflated regression model were used to fit the data.With variable secondary cases and all patients with pleurisy as the response variable and other variables as covariates,including the age of the patient,the number of students of doctor ratio,infectious exposure time,school grades,cases sputum state,PPD strong positive rate,dormitory density degree,levels of ventilation.Results showed that the count model is superior to the traditional model of zero expansion.Through comparison of zero inflated generalized Poisson model fitting effect is best,zero inflated Poisson model and Poisson model fitting effect general,and zero inflated negative binomial model and negative binomial model fitting effect is not good.In influencing factors,tuberculosis continued cases,and the sputum state,PPD strong positive rate and infectious source of exposure time in the model is very significant.In addition,there are dormitories intensive degree,dormitory ventilation degree are more significant.Through the analysis of this paper,it shows that the zero inflation model is more suitable for the analysis of the influence factors and the data fitting of the infectious diseases like tuberculosis.