Prediction of Weight Range of Neonate Using Machine Learning Approach

The weight of a neonate at birth is closely linked to mortality risk in the first year and developmental issues in infancy, and the chance of developing numerous diseases in adulthood. In epidemiological studies, birth weight is frequently considered the causal pathway to these health outcomes. Not all people have the facility to scan their fetus’s weight before birth. In our research, we proposed a method to predict the weight class of the newborn baby by using existing data of pregnant women for the first six months. We identified 16 factors after an extreme literature review and expert opinion. Gradivity, age of pregnant woman, blood group, pregnant woman’s employment, history of abortion, consanguinity, etc., are some of the factors identified. A dataset of 500 pregnant women was collected and preprocessed for applying the machine learning models. Once the preprocessing was done, the dataset was processed with Machine Learning (ML) models. After completion of data preparation, ML algorithms were used for the training and testing data set. This research utilized five ML algorithms to predict newborn baby weight at the six months of pregnancy. ANN and NB models have given accuracy above 70%, SVM and logistic regression algorithms provide accuracy above 60%, and Decision Tree provides accuracy below 60%.