Combination of XGBoost Analysis and Rule-Based Method for Intrapartum Cardiotocograph Classification

The two major components of a cardiotocograph (CTG) are uterine contraction (UC) and fetal heart rate (FHR) signals. CTG has been widely used to monitor well-being in the past 50 years. guideline provided by National Institute Child Health Human Development (NICHD) classifies patterns into three categories (I, II, III) evaluating status fetus. However, manual interpretation is time-consuming subjected inter-personal bias. In this study, we combined rule-based method eXtreme Gradient Boosting (XGBoost) analysis classifying patterns. Because persistent controversies about Category II NICHD, XGBoost was classify it IIa IIb. A total 68 pregnant women were enrolled study. Three consistent both clinicians our algorithm across all categories, average Kappa 0.72. probability distress (FD) 28.8% 71.2% IIb, respectively. These findings show proposed potential provide clinical assistant tool high be an assistive warning system reduce burden medical staff.