THE INTEGRATION OF SUPPORT VECTOR MACHINES AND LOGISTIC REGRESSION FOR IMPROVED DIABETES PREDICTION

Abstract

The increasing prevalence of Diabetes is a global health concern, particularly in Sub Saharan Africa, where the mortality rates are alarming. Projected statistics indicate that by 2045, approximately 12.2% of the world population will be affected by Diabetes. This distressing reality demands collective efforts to combat the disease. Recognizing this, as a manufacturer, I identified an opportunity to address this health crisis while contributing to humanity's well-being. We collaborated with traditional herbalists in Kano State, Nigeria, to obtain local formulations used in the treatment of Diabetes. Subsequently, we conducted a detailed analysis in China, identifying two effective formulations out of nine samples. However, as a non-medical practitioner and a student of Computer Science, I contemplated how I could contribute to this field. A feasibility study revealed that machine learning, known for its prowess in making complex predictions, could be harnessed for Diabetes prediction. This approach would empower doctors to detect the disease at an early stage with heightened accuracy, offering patients a better chance of managing the condition effectively before it becomes lifethreatening.