Classification of Fetal State using Machine Learning Models

¹ Sidi Mohammed Ben Abdellah University, Heigh Normal School Fez, Morocco
² Sidi Mohammed Ben Abdellah University, Doctor Pediatrics Resident at CHU Hassan II, Fez, Morocco
³ Mohammed Premier University, Higher School of Technology, Nador, Morocco

Abstract

In gynecology, the problem of fetus during pregnancy in pregnant women have more interests. In the literature, several means are used to follow the pregnancy such as cardiotocography to measure heart rate, accelerations, fetal movements, and uterine contractions. In this proposed study, we use some algorithms to classify some diseases, and confusion matrix to specify the normal, and suspicious pathology using Random Forest, Support Vector Machine, and Artificial Neural Network. To validate this experimentation, the dataset of UCI has suggested to classify the fetus into three classes: normal, suspicious, and pathological the best performing model for detecting the fetal state is the ANN model which gave better accuracy values for 99.19% for training accuracy and 99.09% for test accuracy.