Study Applicable for Multi-Linear Regression Analysis and Logistic Regression Analysis

ABSTRACT

Current study focus on using method of multi-linear regression analysis and logistic regression analysis, and discuss about the condition and scope of multi-linear regression analysis and logistic regression analysis. A modeling method has been introduced keeping in the basic principles of multi-linear regression analysis and logistic regression analysis. The modeling method and two forms of analytic methods have been analyzed, based on two clinic test data of diabetes and Model-2 diabetes as objects of study in combination with the analytic methods of multi-linear regression and logistic regression. Analysis result indicate that glycosylated hemoglobin, glycerin trilaurate, total cholesterol of serum and blood sugar concentration present obvious positive relation (P < 0.05), whereas insulin and blood sugar present negative relation(P < 0.05); body mass index (BMI) and relative factors are dangerous; physical excise and relative factors are protective. In conclusion, multi-linear regression analysis and logistic regression analysis respectively have their own emphasis; for example, multi-linear regression analysis emphasizes on analyzing linear dependent relation with an dependent variable and multiple independent variables, whereas logistic regression analysis emphasizes on analyzing the relation between probability of occurring an incident and independent variables.

Keywords:

Body mass index (BMI), diabetes, glycosylated hemoglobin, logistic regression analysis, model-2 diabetes, modeling method, multi-linear regression analysis.