Fig. 2

Scheme of our study. Health assessment data, SNP array data and fetal ultrasonography data were used as input of models. The values of fetal ultrasonography data in 26 to 40 weeks of gestation were predicted by AI to improve performance of models. We built the preterm LBW model and term LBW model which predict LBW in preterm birth and term birth group, respectively. The developed models were analyzed to identify influential factors of models and their differences between preterm and term LBW models