Abstract:
The purpose of the study was to perform prediction of maternal outcome in Tanzania using
Machine Learning Classifiers encompassing of Logistic regression, Support Vector Machine
(SVM), Artificial Neural Network (ANN) and Random Forest (RFF). The study employed a
retrospective cross-sectional research design and collected secondary binary data covering a
period of three years from 2020 to 2022 whereby a total of 76,227 women were recorded in The
Health Information Management System (HIMS), so the sample of the study.
The finding suggested that maternal mortality is likely to decrease in future because the survival
cases are predicted more with little evidence of death cases. On the other hand, the study
logistic regression, Artificial Neural Network, and Support Vector Machine predicted maternal
outcome accurately by 99.36%, 93.36%, and 99.96% respectively while, Random Forest
predicted accurately by 100.0%. In that regard, the random forest model and its derived
mathematical equation is suggested ad the best model for predicting maternal outcome in
Tanzania.
The study underscores the complexity of maternal mortality and encourages the development
of targeted interventions and predictive models tailored to the Tanzanian context. Also, the study
recommended that cautions should be taken when suggesting the machine learning predictive
models to avoid overfitting.
