Improved Early Fetal Health Prediction Model using Multi-layer Perceptron: Deep Learning Approach

In the modern era, fetal health risks has evolved as a critical public health concern, indicating the increase of complications in pregnancy, where fetal mortality is the primary risk concern majorly seen in recent years. It effects well beyond the immediate loss, impacting major risks such as preterm birth, neonatal mortality, abnormal growth, long-term health issues. Fetal health monitoring is crucial to ensure safe pregnancy results and to identify early complications. Early prediction of fetal health can minimize the complications during pregnancy, and results positive outcome. If early diagnosis is neglected, it may cause long-term consequences. Here the Cardiotocography (CTG) is the essential technique in obstetrics to monitor fetal health, it obtains necessary information to ensure the well-being of fetus. This study presents a machine learning based approach for early prediction of fetal health risk based on some primary factors. We propose the Multilayer Perceptron(MLP) model that is applied on the dataset to train the model by dividing the dataset into training and testing sets. The model trained with data on Fetal health was simulated using the Waikato Environment for Knowledge Analysis (WEKA) tool. The prediction of fetal health is carried out based on three categories: Normal, Suspicious, and Pathological. The proposed method achieves the highest accuracy of $99.1 \%$ compared to other models.