Fetal health prediction based on Adaptive Grey Wolf Optimizer - BP Neural Network (AGWO-BPNN)

Fetal health prediction is an essential approach to reduce perinatal and neonatal mortality, and BP Neural Networks have emerged as a research hotspot in the field of medical big data due to their tremendous nonlinear mapping capacity, but there are constraints in training efficiency and predictive accuracy. This paper proposes a BP Neural Network based on Adaptive Grey Wolf Optimizer to improve the accuracy of predicting fetal cardiogram (CTG) health status. When compared to the baseline BPNN and the classical GWO-BPNN, AGWO-BPNN outperforms it in terms of accuracy, recall, AUC value, and F1 score, indicating that the Adaptive Grey Wolf Optimizer (AGWO) is beneficial in improving the prediction performance of BP Neural Networks. The AGWO-BPNN model is a valuable tool for clinical decision support systems.