Fetal health classification from cardiotocographic data using machine learning

TLDR

Health complications during gestation are a global issue, often causing fetal mortality especially in developing countries, and machine learning has advanced diagnosis, treatment, and prognosis in healthcare. The study aims to predict fetal health from cardiotocographic data by labeling states as normal, needs guarantee, or pathology using ML algorithms such as SVM, RF, MLP, and KNN, and proposes that feature engineering could further improve results. The authors employed regression and correlation analyses together with ML classifiers (SVM, RF, MLP, KNN) to evaluate how CTG attributes influence fetal health. Random forest outperformed the other algorithms in accuracy, precision, recall, F1‑score, and support.

Abstract

Abstract Health complications during the gestation period have evolved as a global issue. These complications sometimes result in the mortality of the fetus, which is more prevalent in developing and underdeveloped countries. The genesis of machine learning (ML) algorithms in the healthcare domain have brought remarkable progress in disease diagnosis, treatment, and prognosis. This research deploys various ML algorithms to predict fetal health from the cardiotocographic (CTG) data by labelling the health state into normal, needs guarantee, and pathology. This work assesses the influence of various factors measured through CTG to predict the health state of the fetus through algorithms like support vector machine, random forest (RF), multi‐layer perceptron, and K‐nearest neighbours. In addition to this, the regression analysis and correlation analysis revealed the influence of the attributes on fetal health. The results of the algorithms show that RF performs better than its peers in terms of accuracy, precision, recall, F1‐score, and support. This work can further enhance more promising results by performing suitable feature engineering in the CTG data.