Predictive System for Fetal Health and Pregnancy Risk

Health assessment of the fetus and monitoring the possible risks in the pregnancy are crucial for safety of the mothers and lowering complications. Methods based on CTG and ultrasound use human interpretation that might cause mistakes due to insufficient resources. This research proposed using an artificial intelligence model to increase accuracy and time-efficient classification of fetal health and pregnancy risks. The machine learning model classifies risk into low, average and high according to the data about the mother’s and baby’s health for taking necessary measures. The following models will be applied in this research – SVM, K-Nearest Neighbors, Decision Tree, Random Forest, ANN, AdaBoost, Gradient boosting, XGBoost and Multilayer Perceptron due to their ability to classify the data. The result of the tests proves that ensemble methods and neural networks and more accurate and reliable. AdaBoost and multilayer perceptron showed better performance compared to other methods.