Prediction method and system for preeclampsia risk of early pregnancy

The invention provides a prediction method and system for preeclampsia risk of early pregnancy, and relates to the technical field of noninvasive prenatal detection.The method comprises the steps that firstly, sample measured value data of the same pregnant day are input into a preset kmeans unsupervised learning model, the sample measured value data are preprocessed, and parameters of k clusters are obtained; secondly, on the basis of parameters of each cluster, searching and comparing the sample size of each cluster to find out low-risk samples; then calculating the median multiple of the sample measurement value data of each pregnant day in the low-risk sample and a difference value corresponding to the median multiple of the evaluation parameter of the early pregnancy eclampsia early-stage risk model; and optimizing and updating the early-stage pregnancy preeclampsia risk model by using a difference value so as to predict the early-stage pregnancy preeclampsia risk. According to the method, low-risk sampl