Comparative Analysis of Machine Learning Models For Early Detection of Fetal Disease using Feature Extraction

In today’s world, fetal diseases has become a common problem faced by the majority of pregnant woman around the globe. The fetal disease is a serious ingrown malformation in children. These diseases can be diagnosed through heart functioning which is identified using Fetal Electrocardiogram signals. Early detection of such fatal diseases can save many lives. To this end, echocardiogram signals remain the most effective method and our experimental research is quite relevant. This research work is an involved comparative study of different classifier models for the early diagnosis of fetal disease. The experiment consists of 10 classifier models including sequential models with hyper-parameter tunning. Each classifier model is evaluated based on its accuracy, recall, precision, and other evaluation parameters. The experimental results show that XGBoost and LightGBM Classifier models show accuracy as high as 95.14% followed by the Gradient Boost classifier model with an accuracy of 94.67%. However, the maximum recall value was achieved by the XGBoost classifier.