Enhancing Fetal Health Classification: A Study on Fetal Cardiotocograms

Ensuring the well-being of the fetus during pregnancy is crucial for preventing adverse outcomes in childbirth. This paper addresses the imperative need for accurate fetal health Classification through the utilization of advanced machine learning (ML) and deep learning (DL) techniques. The dataset employed comprises 2126 records extracted from UCI Machine Learning Repository fetal cardiotocograms (CTGs) exams with features encompassing data related to fetal heart rate (FHR) and uterine contraction (UC) gives intra uterine pressure, categorized into three classes: normal, suspect, and pathological by expert obstetricians. Our study leverages a combination of ML and DL algorithms to enhance fetal health Classification. Traditional ML algorithms, including k-NN, DT, SVM, RF, SGD, GB, AdaBoost, and XGBoost, demonstrate high accuracy ranging from 97.37% to 99.50%. Additionally, the application of DL methods, such as convolutional neural network, Long Short-term memory, and Bi-directional Long Short-term memory, yields notable accuracies of 98.20%, 99.37%, and 98.74%, respectively.