Application of Machine Learning in Assignment of Child Delivery Service in Afghanistan

Skilled assistance during child delivery is essential to reduce maternal mortality. However, in Afghanistan, which has the highest rank of maternal mortality globally, the coverage of birth attended by skilled personnel remains very low. Therefore, the ability to assign skilled child delivery service is beneficial for efficient use of personnel resources and a good preventive measure. This study attempts to implement a predictive model using data mining classification algorithms to determine whether a skilled child delivery service is necessary. The study also attempts to identify the most suitable classifier among the five popular machine learning techniques. These are Random Forest (RF), Neural Network (NN), J48, CART rule induction and Support Vector Machine (SVM). The dataset used is the Afghanistan Demographic and Health Survey (AfDHS). The classification in this study is binary, whether a skilled child delivery service is `necessary' or `not necessary'. The Correlation-based Feature Selection (CFS) algorithm is adopted as a feature selection method, where accuracy, precision, recall, and area under the ROC curve (AUC) are used as evaluation metrics. The result of this study ought to be beneficial to the Afghanistan Health Care Sector.