Fetal Health Status Classification Using MOGA - CD Based Feature Selection Approach

Pregnancy is a cycle that invites us to give up on the hidden power behind all life. A stable pregnancy is a good thing. The likelihood of safe pregnancy is increased by timely and daily fetal treatment. Fetal welfare is a program that controls a pregnant woman during her pregnancy. Every pregnant woman should know what the result of pregnancy is about. Every day about 800 women die from pregnancy and birth complications. Maternal and fetal health have a close relation, as nearly three million newborns die every year. Therefore, adequate treatment is important, including risk evaluations for the health of both mother and child prior, after and during delivery. CTG (Cardiotocography) is a way to monitor and predict fetal well-being in females with a complex issue. Through supplying data on infant heart rate from the abdomen of the mother, CTG avoids premature birth. In this research, we focused on an evolutionary multi-objective genetic algorithm (MOGA) for extracting important factors causing fetal death by cardiotocographic analysis of fetal evaluation. In particular, the feature importance value is taken into consideration when there is a tie among the set of non dominated solutions. Seven existing classification models (LR, SVM, RF, DT, KNN, GNB, XGBoost) are used to test its efficiency concerning fetal health classification in the dataset of most relevant features.