bayesian averaging

Diabetes mellitus is one of the most common metabolic disorders in the world. It has many risk factors, and the prediction of these factors is important to prevent it. The current study aimed at determining the factors associated with the incidence of type 2 diabetes using the Bayesian regression model of spike and slab.
In the current descriptive study, data from 819 participants in the diabetes screening program at Zahedan Health Centers, Iran in 2014 were analyzed. The convenience sampling method wasused. The demographic information was collected from patient's records, including gender, age, history of blood pressure,ect., and fasting blood glucose level. For prediction and variable selection, the Bayesian regression model of spike and slab was used, then validity of the model was assessed by mutual validation method. Data analysis was performed with R software version 3.3.2.
The family history of diabetes and age were significantly correlted with diabetes in the subjects;the family history of diabetes with a 100% reliability coefficient was the most effective variable.
The Bayesian regression model has a good ability to dignose diabetes; while recognizing less effective variables, the predicting ability of the model was also maintained. Given the factors associated with the incidence of diabetes, the implementation of screening plans with priority given to the elderly and those with a family history of diabetes is suggested to control the outbreak of diabetes.