Predictive Supervised Machine Learning Models for Diabetes Mellitus.

Area of Science:

• Medical Informatics
• Public Health
• Machine Learning

Background:

• Diabetes mellitus (DM) is a global health crisis, increasingly prevalent in developing nations like Nigeria.
• Type 2 DM constitutes approximately 90% of all DM cases, with significant projected increases worldwide.
• DM poses a growing public health concern in Nigeria, necessitating advanced diagnostic tools.

Purpose of the Study:

• To develop and evaluate supervised machine learning models for predicting type 2 diabetes.
• To identify the most effective machine learning algorithms for DM diagnosis in the Nigerian context.

Main Methods:

• A diagnostic dataset for type 2 DM was collected from Murtala Mohammed Specialist Hospital, Kano.
• Supervised machine learning models were developed using logistic regression, support vector machine, K-nearest neighbor, random forest, naive Bayes, and gradient boosting algorithms.
• Model performance was evaluated based on accuracy and receiver operating characteristic (ROC) curve analysis.

Main Results:

• The random forest model achieved the highest accuracy at 88.76%.
• Both random forest and gradient boosting models demonstrated the best predictive ability with 86.28% using ROC curves.
• The study successfully applied various machine learning techniques to a real-world diabetes dataset.

Conclusions:

• Machine learning models show significant potential for accurate type 2 diabetes prediction.
• Random forest and gradient boosting algorithms are particularly promising for DM diagnosis in Nigeria.
• Further research and implementation of these models can aid in early detection and management of diabetes.