Fig. 5: Partial dependence plots for the eight features selected by Boruta and used in the random forests classifier.
From: Language function following preterm birth: prediction using machine learning
a The predicted language impairment probability rises with increasing PSFA values. b 3D plot of PSRD and PSAD. The predicted language impairment probability rises with increasing PSRD and PSAD values. c A twin pregnancy increases the predicted probability of language impairment. d An incomplete course of antenatal corticosteroids increases the predicted probability of language impairment. e No exposure to any antenatal steroids increases the predicted probability of language impairment. f Female sex reduces the predicted probability of language impairment. g Feeding with exclusive breast milk reduce the predicted probability of language impairment. Language composite score <85 at 2 years CGA is more likely following a twin pregnancy, an incomplete course of antenatal corticosteroids, or no exposure to any antenatal steroids. Female sex and feeding with exclusive breast milk reduce the risk of future language delay.
