Fig. 7
From: Fetal weight estimation based on deep neural network: a retrospective observational study
A-D The performance of models after feature selection in the primiparas and multiparas groups and the entire cohort. The mean deviation (± 1 SD), (A) mean absolute error (MAE), (B) mean absolute percentage error (MAPE), (C) and root-mean-square error (RMSE), (D) and their 95% confidence interval of the EFW of multiparas group and the entire cohort after model-free Greedy (MFG) algorythm feature selection versus Hadlock’s formula and the all features model