Marginal Regression Analysis with Two Basic Parameterizations and the Dependency Ratio of Maximum Likelihood Estimation
Pages : 921-926Download PDF
In this paper we have developed an Marginal Regression for a bi-variate response for untreated state, diabetes mellitus is recognized by chronic evolution of concentration of glucose in the blood (Hyperglycaemia). This is sometimes accompanied by symptoms of serve thirst, profuse urination, weight loss, and stupor culminating in coma and death in the absence of effective treatment. The underlying caves of diabetes are the defective production or action of the hormone insulin. We are tested through SPSS software by taking 200 samples and we determined this model by care processing, dependent variable encoding classification table Omnibus tests of model coefficients, and we also developed a model summary, Model if term removed variables in the equation and variables not in the equation. From above study we notice that explanatory variables age and DH are the significant variable as the before.
Keywords: Marginal Regression Analysis,Binary Logistic Regression, logit model, Odds Ratio, Model validation, Hosmer and Lemeshow Test.