Study of Binary Logistic and Poisson Regression Models of Diabetic Patients in Nigeria using Dichotomous and Non- Dichotomous Predictors
Onu, Obineke Henry *
Mathematics and Statistics Department, Ignatius Ajauru University of Education, Rumuolumeni, Port Harcourt, Rivers State, Nigeria.
Amakuro, Okuata Avula
Mathematics Department, Isaac Jasper Boro College of Education, Sagbama, Bayelsa State, Nigeria.
Alabge, Samson Adekola
Mathematics Department, Isaac Jasper Boro College of Education, Sagbama, Bayelsa State, Nigeria.
*Author to whom correspondence should be addressed.
Abstract
The comparative study of the Binary-logistic and Poisson regression models of diabetic patients in Nigeria was presented using R-squared, Adjusted R-squared, Variance Inflated Factors and Akaike Information Criterion for two different data sets of Diabetic Patients known as the dichotomous and the Non-dichotomous data obtained from the University of Port Harcourt Teaching Hospital (UPTH). The results revealed that the Binary logistic regression was better than the Poisson regression for both dichotomous and non-dichotomous data. It was also, observed that, the Binary-logistic regression model was significant in this study with a non-dichotomous data set, while Poisson regression was not significant. The results also showed that both the type 1 and type 2 diabetes have negative effects on the diabetic Patients.
Keywords: Binary-logistic, poisson regression, dichotomous data, non-dichotomous data, diabetes type 1 diabetes type 2 diabetes