MACHINE LEARNING FOR PREDICTING PREECLAMPSIA BASED ON ARTERIAL PRESSURE AND BODY MASS DATA OF THE PREGNANT
Abstract
This article examines the task of early prediction of preeclampsia using machine learning methods based on data from pregnant women's blood pressure and body weight. The work describes a scheme for collecting clinical samples, preliminary data processing, and constructing a classification model based on a random forest algorithm. It has been shown that even simple, widely available indicators - systolic and diastolic blood pressure, blood pressure dynamics, body mass index, and weight gain - allow for high discrimination of the model, with good sensitivity and moderate specificity. The results obtained confirm the prospects for using non-invasive, easily measurable parameters and machine learning methods to support clinical decision-making in obstetric practice.