Using TensorFlow to Establish multivariable linear regression model to Predict Gestational Diabetes

Gestational diabetes mellitus (GDM) can increase the risk of fetal distress or stillbirth, which seriously affects the health of mothers and infants. Therefore, it is important to strengthen the prediction of GDM in early pregnancy to reduce the occurrence of GDM[1]. Through collecting the data of obstetrics and gynecology clinic in Chifeng Hongshan District Maternal and Child Health Care Center. After data cleaning and pretreatment, a total of 419 pregnancy data sets were collected. The body mass index (BMI), the pre-pregnancy BMI, the maternal age and the family history of diabetes, and the fetal abdominal circumference on the day of pregnancy were selected as independent variables, and a multivariable linear regression model was used to learn. In the predictive model of diabetes, the least squares method and stochastic gradient descent algorithm were used to optimize the training and the prediction model was evaluated. The random error was 3.2 and the loss rate was 0.02.