Cardiotocography’s Use of ML Models for Classifying the Health of a Foetus

Cardiotocography (CTG) being valuable tool for tracking the fetus’s health by giving vital information on the FHR(fetal heart rate) and UC (uterine contractions). In Our work, we present ML-based method for dividing fetal health into groups of three: Suspect, Pathological and normal is defined by CTG data. Utilizing a variation of ML approaches, including Support Vector Machines (SVM), Gradient Boosting, DT (Decision Trees), and RF (Random Forest), we generate an eternal classification model. Significant CTG statistics, such as baseline heart rate, accelerations, decelerations, and uterine contractions, are included in the dataset that was utilized. We employ stratified k-fold cross-validation to ensure balanced training and consistent significances. Along with ROC-AUC, other metrics like F1- score, recall, accuracy and precisions is also utilized in the assessing model. This method seeks to progress prenatal care and lower fetal risks while giving medical professionals prompt diagnosis and response.