Fetal Health Classification Using Machine Learning on Cardiotocography Data

This study explores a machine learning approach for detecting and classifying fetal health which is a crucial aspect in ensuring proper care and wellbeing of both fetus and the mother. The conventional diagnosis methods were manually interpreted from cardiotocography data which leads to subjective or erroneous decisions. With the help of Machine learning and accessing cardiotocography data (CTG) it provides some crucial advantages for identification of different fetal physiological states (Normal, Suspected and Pathological) by analyzing different fetal heart rate patterns and provide accurate observations. Analyzing various health parameters, including physiological indicators and factors a elaborate dataset has been utilized to train and evaluate supervised machine learning models. KNN, random forest, logistic regression, decision trees, neural network and gradient boosting has been employed to optimize predictive accuracy. The paper encompasses rigorous preprocessing steps, machine learning algorithm performance and evaluation metrics. Results demonstrate the effectiveness in early detection, providing a foundation for proactive healthcare, diagnostic accuracy, improves maternal-fetal outcomes while addressing issues such as data imbalance and interpretability. Future advancement focuses real time analysis, IOT integration which will bridge the gap between clinical implementation and adoption.