Predictive Model for Early Detection of Mother's Mode of Delivery with Feature Selection

Abstract

At childbirth, a decision needs to be taken regarding the most suitable mode of delivery for mothers. Often, certain historical factors account for this decision, some of which are based on individuals’ personal choices. In this study, secondary data containing attributes and mode of child delivery were analyzed to predict the mother’s mode of delivery using machine learning techniques. We built a predictive model with four different machine learning algorithms where a recursive feature elimination technique was employed to rank the most important feature attributes. Our study shows that mother’s Length of Stay, their Number of Visits to the hospital, and the Number of Assisted Delivery Procedures emerged as the most important attributes for predicting the mode of delivery while Parity, Educational Level, and Location (residence) were the least important. We envision that these findings will guide policy and practitioners’ decisions toward the mode of child delivery of women in Nigeria. Target Audience This book chapter targets medical and healthcare professionals and practitioners, especially those associated with maternal and newborn health and pregnant women. The chapter seeks to guide the choice of delivery mode for expectant mothers to help reduce the rate of morbidity and mortality associated with childbirth. In making choices for the mode of delivery, length of stay, number of visits to hospital, and number of previous cesarean procedures on the expectant mother were found to be critical determinants. Thus, early detection and prediction of the right delivery mode will help avert possible complications, prevent, or reduce both maternal and child mortality.