The invention relates to the technical field of
medical diagnosis, and discloses a fetal
nervous system auxiliary diagnosis method based on
big data. The method comprises the following steps: collecting
nervous system monitoring data of a plurality of fetuses; performing prenatal whole
pregnancy dynamic tracking regularization
processing on the ultrasonic
radio frequency signal sequence, aligning ultrasonic characteristic differences among different fetuses, and generating a fetal
nervous system development ultrasonic characteristic dynamic change curve; obtaining a non-
pathological feature weight and an optimized association degree quantity of each sampling point, calculating a comprehensive importance
score of each sampling point based on a product of the non-
pathological feature weight and the optimized association degree quantity, and selecting the sampling points of which the scores are higher than a dynamic threshold value as key feature points; integrating amplitude data and clinical health scores of the regularized feature dynamic change curves of all fetuses at key feature points, and training a diagnosis prediction model by using a
gradient boosting decision tree algorithm; and
processing the ultrasonic
radio frequency signal sequence of the
fetus by using the diagnosis prediction model, extracting the amplitude of the key feature points, inputting the amplitude into the model, and outputting a nervous
system development outcome risk prediction evaluation value.