An Application of Gradient Boost Regressor in Birth Time Prediction

Birth Time to Reduce Risks of Maternal Mortality

Abstract

About 40% of expectant mothers in Kenya give birth at home. In fact, data from the Ministry of Health on maternal mortality show a significant decline in skilled delivery by 1.3 per cent, in particular, 80.6 per cent in 2019 to 79.3 per cent in 2020. Deliveries at home pose high risks of maternal deaths, thus a need to reduce the number of deliveries that occur at home to improve maternal health. The major problem that hinders emergency response when a mother is in labour, especially in remote areas, is distance from the hospital. In West Pokot and Kilifi Counties, most mothers live over 10 km away from the nearest health facility; this is a factor that cuts across most rural setups in Kenya. Thus, the need for prediction. This project employs the Gradient Boost Regressor to make the best possible prediction of a mother's expected day of delivery in order to avail emergency services early to avoid the risks of giving birth at home, preventing maternal mortality.