Enhancing Childbirth Delivery Mode Prediction: A Machine Learning Hybrid Ensemble Approach

Decision-making in obstetrics, like identifying high-risk cases that require timely intervention, can be improved by using machine learning algorithms that predict childbirth modes, vaginal or cesarean. Such an algorithm might consider factors like age and medical history of pregnant women. These algorithms also help healthcare providers to plan and allocate resources to suit specific patients. Using an accurate predictive model with ML algorithms supports informed decision-making in childbirth modes. Recent technological advancements have contributed to reducing maternal mortality rates, yet ensuring the health of both mother and child during pregnancy remains a challenge. Predictive modeling has gained prominence in addressing this challenge, enabling anticipating problems and implementing preventative measures to mitigate pregnancy-related risks. The hybrid technique employed in this study utilizes Naïve Bayes, K Nearest Neighbor, Decision Tree, and Logistic Regression as base classifiers, with Random Forest serving as the meta-classifier. Noteworthy performance metrics include Linear Discriminant Analysis achieving 92% accuracy, Support Vector Machine reaching 92% accuracy on the same dataset, and the proposed hybrid model outperforming with an accuracy of 99%.