Building Statistical Model for Predicting Risk of Diabetes

Abstract:
In recent years, diabetes has become one of the most common human diseases in the world, and is even the main cause of high mortality and economic losses, while timely diagnosis and prediction provide patients with appropriate methods for prevention and treatment. By using a logistic regression model, we tried to predict type 2 diabetes. The statistical analysis was conducted with SPSS for descriptive analysis of data, a chi-square test, and logistic regression analysis to predict the risk factor of diabetes. As the result, five main predictive factors were identified: waist circumference, family history, hypertension, cardiovascular disease, and age. The overall prediction rate of the logistic regression model for predicting diabetes was 80%. The research results help prevent the occurrence of diabetes or facilitate early treatment, reduce misdiagnosis and avoid wasting health care resources.