An Intelligent Fetal Health Diagnosis Framework Leveraging a RVFLN with Feature Selection Techniques

Fetal health is a vital component of prenatal care, and a thorough early examination of it is essential to prevent difficulties including stillbirth, new born discomfort, and long term developmental issues. Cardiotocography (CTG) is widely used by obstetricians to monitor the health of both mother and fetus. An obstetrician’s interpretation of CTG data after visual analysis is quite simple however it might not be as precise. Because of this, the need to classify the fetal state from CTG recordings has prompted a number of studies to employ new computational intelligence techniques. This study proposes a predictive framework utilizing a random vector functional link neural network (RVFLN) for fetal health classification. To enhance its predictive performance, three Feature Selection (FS) techniques such as Mutual Information (MI), ReliefF and Analysis of Variance (ANOVA), are employed to extract the most significant features from the CTG data. Comparing the performance of proposed RVFLN model with three FS techniques, it is evident from the experimental findings that it achieves an accuracy of 91.5% with a minimum of 5 features chosen by MI-based FS technique. The results highlight the potential of the proposed approach in providing a reliable and efficient solution for fetal health monitoring and risk prediction.