Predicting High-Risk Perinatal Complication Using Semi-supervised Machine Learning

Machine learning algorithms are employed to analyze medical history data and detect potential pregnancy complications by examining data linked to a pregnant woman's medical history. These algorithms possess the ability to discern patterns within the data that might indicate an elevated likelihood of specific pregnancy complications. The primary objective of this proposed approach is to identify unforeseen pregnancy complications. Through the application of machine learning methodologies, healthcare professionals can enhance their ability to recognize and address pregnancy-related issues more efficiently. The proposed approach employs a variety of supervised machine learning algorithms to attain the highest achievable accuracy score for predicting the elevated risk of preterm labor complications. To bridge existing gaps, the suggested self-training decision tree incorporates a 30% dataset that lacks labeling. Remarkably, this approach yields accuracy surpassing other algorithms, reaching a maximum accuracy of 89%.