Investigating Correlations Between Maternal Mortality Rate and National Developmental Status

According to the World Health Organization (WHO), 287,000 women were victims of maternal mortality in 2020, and this rate has been increasing at an alarming pace since then. However, the likelihood of experiencing maternal mortality is heavily influenced by socioeconomic factors, including a person's place of residence and the level of development in their country. We hypothesize that incorporating AI and machine learning techniques can help uncover correlations between a country's development and its maternal mortality rates.

In this study, data from Kaggle was used to develop machine learning models. The dataset includes three key features: the level of human development, whether the country is part of a UNDP developing region, and its Human Development Index (HDI) rank. The dataset also provides maternal mortality rates per 100,000 births from 1990 to 2021. We analyzed 195 data points from various countries, and with the help of the IBM platform, we were able to apply multiple algorithms to develop predictive models.

The root mean squared error (RMSE) values of the models were found to be relatively high, exceeding 200. This is currently under further investigation to improve the accuracy of the models by addressing the issues related to imbalance nature of data being employed in this study. Once this issue is resolved, we will examine the significant features contributing to maternal mortality and identify emerging trends. To aid in the analysis, we obtained ROC curves, F1 scores, and confusion matrices for a comparative discussion. A detailed discussion of the machine learning results, along with strategies to improve model accuracy, will be presented.
In the future, this model will be expanded to include a larger dataset from various sources. The goal is to explore the potential of machine learning in developing predictive models that can reveal intra-national trends, ultimately saving lives.