MODEL UNTUK DATA BERDISTRIBUSI POISSON YANG MENGANDUNG VARIABEL BEBAS KUALITATIF

Abstract

The main purpose of this research is to build model of data that distribution of respon variable is Poisson and dependent variable has qualitative variable, beside that, to know estimation of parameters and testing of hipotesis. This research uses Generalized Linear Model. At its analysis, qualitative variable is considered according to probability regression function of categories with dummy variable. Formula that can be used is Y~G*X to know if the categories need apart or not. Then to know explicitly how influence intercept and slope of each categories can be used Y~G/X-1 formula. From both of analysis, then doing selection which the best model. Selection of the best model also can be look at from AIC(Akaike’s Information Criteria). The better model has AIC smallest.