Wearable Sensors and AI Algorithms for Monitoring Maternal Health

2. Artificial Intelligence For Pregnant Women Healthcare

Pregnancy outcomes can be detected with the aid of AI and ML methodologies, including contemporary deep learning techniques. During prenatal treatment, accurate diagnosis and prediction techniques can aid in the earliest potential problem detection. A supervised artificial neural network (ANN) (Beksac et al., 2018) has been created in order to forecast the delivery path. Gestational age at birth, maternal age, risk factors, and disorders of the mother are the variables used as input for ANN. The programme outputs two options: vaginal delivery or caesarean section (CS) (VI). Data from the patients who were admitted for delivery and who were chosen at random to train the algorithm were used for the experiment. According to the presented findings, the ANN-based system has an efficiency rate of 97% and outperforms other statistical tests.

In a different study, a prediction technique was put up to forecast foetal congenital abnormalities (Akbulut et al., 2018). To train and process the data to predict the foetal abnormality status, a variety of binary classification models, including SVM, decision forest, neural network, decision jungle, and others, were used. RadyoEmar radio diagnostics centre in Istanbul, Turkey provided the information via a questionnaire, and an e-health application was created to use the parameters as input. The Decision Forest algorithm, which was tested, produced the best predictions in terms of accuracy, AUC, and F1 score when the data was trained with it.