Identification of Risk Factors and Development of Machine-Learning-Based Forecasting Models for Infant and Maternal Mortality in Bangladesh

Maternal Mortality Rate (MMR) and Infant Mortality Rate (IMR) are key indicators of health and socioeconomic conditions in developing countries, with Bangladesh as a prominent case. This study aims to identify risk factors associated with maternal and infant mortality in Bangladesh and to construct prediction models for MMR and IMR using machine learning techniques. It also investigates common features and similarities between MMR and IMR, which are closely related and often move together over time. We analyze annual data for Bangladesh from 1991 to 2017, sourced from the World Bank’s 2021 World Development Indicators (WDI), allowing examination of long-term patterns rather than short-term fluctuations. Statistical analysis focuses on relationships between mortality indicators and their associated factors. All factors retained in the final models are statistically significant (p ≤ 0.05). The prediction models achieve R-squared values very close to one, indicating that the selected factors explain almost all variation in the mortality indicators. Support Vector Regression (SVR) and stepwise Linear Regression (LR) are applied to select influential variables, estimate their contributions, and generate predicted values for MMR and IMR. Model performance is evaluated using Root Mean Square Error (RMSE), Mean Square Error (MSE), and Mean Absolute Error (MAE), with LR and SVR outperforming other approaches. The efficiency and accuracy of these models provide useful tools for understanding mortality patterns in Bangladesh, examining how changes in key risk factors are associated with MMR and IMR over time, and informing discussions on reducing maternal and infant mortality in the country.