Dengue Haemorrhagic Fever (DHF) is a disease caused by the dengue virus is transmitted through the bite of the mosquito Aedes aegypti In 1994 cases of dengue fever began to spread to 27 provinces in Indonesia, including the province of Central Java in particular Tegal. Tegal District Health Office noted the number of dengue cases in Tegal regency in 2014 reached 526 cases.This analysis aims to determine the factors that influence prevention DHF in Tegal regency. The dependent variable in this study is the number of dengue cases in 2014 while the independent variable is the number of health centers, health workers, population, implementation of fumigation (fogging), and household clean and healthy lifestyle. The analysis used in this research is descriptive analysis, Poisson regression and negative binomial regression. Before performing poisson regression analysis is the assumption that must be done is multicolinierity test to determine whether or not the relationship between the independent variables. Poisson regression is a nonlinear regression that is often used to model the response variable in the form of natural numbers. Poisson regression model has equidispersi assumption, namely that the mean and variance of the response variable of the same value. Results of regression analysis poisson This is a variable number of health centers, health workers, population, implementation of fumigation (fogging), and household pattern of clean and healthy (PHBs) affects DHF, but in fact a violation of the assumptions in the regression poisson namely the overdispersi (variance value is greater than the value of the mean) so that the Poisson regression model is not appropriate for use in this study. Then the appropriate measures to overcome the overdispersi by using negative binomial regression. The results of negative binomial regression analysis were derived variable implementation of fumigation (fogging) effect on dengue fever in Tegal regency.
Keywords : multikolinieritas, overdispertion, Poisson regression, negative binomial regression.