STUDY THE EFFICIENCY OF JEAN BOURGEOIS PICHAT MODEL TO PREDICT INFANT MORTALITY COMPARED TO MACHINE LEARNING RANDOM FOREST

Data on infant mortality is essential for demographic analyses and is a critical metric for gauging a nation's socioeconomic development and standard of living. It also aids in identifying kids who are in danger of passing away and creating strategies to reduce such risks, like encouraging birth spacing. So far, no research has been done in the Gulf Cooperation Council (GCC) region that has used machine learning (ML) techniques; nevertheless, prior studies have mainly relied on conventional regression analysis. To estimate newborn mortality rates, this study used ML techniques to evaluate the Jean Bourgeois Pichat (JBP) model's predictive power to other ML models. Data from the National Center for Statistical and Information (NCSI) were subjected to ML technique of Random Forest using Python Software 3.9. The performance of the model was evaluated using metrics such the confusion matrix, precision, accuracy, F1 score, recall, and AUROC. The JBP model was used to forecast the number of newborn deaths, and the Chi-square statistical test for independence was used to examine a range of variables, including socioeconomic, communal, bio-demographic, and environmental factors. The Random Forest model has the following statistics: 85.2% accuracy, 80.3% precision, 81.2% F1 score, 90.3% recall, and 85.2% AUROC. The number of newborn fatalities may be accurately predicted using the JBP model. Overall, the results of the study showed that the Jean Bourgeois Pichat (JBP) model can predict infant mortality better than the machine learning (ML) random forest model.