Medical data is being generated on a large scale and so, there is a need to build models that can predict diseases accurately to assist medical practitioners in the early detection of chronic diseases such as diabetes. Machine Learning models are currently being developed; however, their accuracy must be enhanced to reduce false negatives. The effect of false negatives can be devastating as a likely diabetic might be predicted as nondiabetic. Diabetes is a disease that affects people around the world. Patients with diabetes are at risk of developing other health complications such as retinopathy, heart disease and kidney failure. The goal of this study is to build a model trained on dataset that is representative of the target population. We proposed a fused machine-learning model consisting of linear SVM (Support Vector Machines) and XGBoost. The proposed model is expected to have better prediction accuracy due to the combination of the robust performance metrics of both SVM and XGBoost.
Author: Kemi Akanbi