Machine learning to predict unintended pregnancy among reproductive-age women in Ethiopia: evidence from EDHS 2016

Abstract

Abstract Background An unintended pregnancy is a pregnancy that is either unwanted or mistimed, such as when it occurs earlier than desired. It is one of the most important issues the public health system is currently facing, and it comes at a significant cost to society both economically and socially. The burden of an undesired pregnancy still weighs heavily on Ethiopia. The purpose of this study was to assess the effectiveness of machine learning algorithms in predicting unintended pregnancy in Ethiopia and to identify the key predictors. Method Machine learning techniques were used in the study to analyze secondary data from the 2016 Ethiopian Demographic and Health Survey. To predict and identify significant determinants of unintended pregnancy using Python software, six machine-learning algorithms were applied to a total sample of 7193 women. The top unplanned pregnancy predictors were chosen using the feature importance technique. The effectiveness of such models was evaluated using sensitivity, specificity, accuracy, and area under the curve. Result The ExtraTrees classifier was chosen as the top machine learning model after various performance evaluations. The region, the ideal number of children, religion, wealth index, age at first sex, husband education, refusal sex, total births, age at first birth, and mother’s educational status are identified as contributing factors in that predict unintended pregnancy. Conclusion The ExtraTrees machine learning model has a better predictive performance for identifying predictors of unintended pregnancies among the chosen algorithms and could improve with better policy decision-making in this area. Using these important features to help direct appropriate policy can significantly increase the chances of mother survival.

Keywords

• Unintended pregnancy
• Machine learning
• EDHS data
• Ethiopia