Estimating the impact of Pregnancy, Systolic and Age on Diabetes for Women by Using Support Vector Regression Model (SVR)

Abstract

In this paper we have 623 cases of diabetes patients, the data partitioned in to training dataset (500 observations) and testing dataset (123 observations), and the aim is to estimate the impact of pregnancy duration in weeks, systolic and age as factors on diabetes of the women patients for this purpose SVR has been used. According to the results radial kernel function gave highest performance compared to the other kernel functions, the R2 = 83% this implies the factors capable of explaining 83% of diabetes variable with MSE and RMSE of (0.000958 and 0.030956) respectively. And p-values of the three aforementioned variables are less than the significant level of 0.01, implying that the three factors have a statistically significant impact on the response variable. Where Pregnancy duration in weeks has an impact of 0.401 on the patient, that means if duration increase by one week, then diabetes will increase by 0.401 units, also both Systolic and age have a significant positive effect on the response variable, and the amount of impact is (0.621 and 0.557) respectively.

Keywords

• Regression Model
• Support vector regression
• kernel functions