A systematic review of machine learning methods for diabetes mellitus prediction and classification in Nigeria

Abstract

Diabetes mellitus (DM) is a significant public health issue in Nigeria, affecting millions of people. Early detection and management are crucial to prevent severe complications. Traditional diagnostic methods have limitations, prompting the exploration of machine learning (ML) techniques for more accurate and efficient prediction. This review systematically examines existing studies on ML applications for DM prediction and classification in Nigeria. It analyzes the research methodologies, attributes considered, study areas, and performance of different ML algorithms. The findings reveal that while ML holds promise, research is limited in scope, focusing primarily on the northern regions. Supervised learning algorithms like ANN and decision trees have demonstrated promising results for prediction and classification in Nigerian datasets, with logistic regression being a common tool for risk factor analysis. Furthermore, studies often overlook key risk factors prevalent in the southern population. This review highlights the need for future research that considers the southern population and a wider range of risk factors. It further recommends a decision support system to improve the early detection, management, and outcomes of diabetes in remote regions of Nigeria.