An Ensemble Learning to Predict Glucose Level of Patient by Urine Sample

Abstract

All areas of our lives are evolving and growing with machine learning (ML) including the healthcare system. In healthcare systems, diabetes is one of the extremely deadly diseases that could cause patients multiple organ failures, if we cannot take significant action on time. In this paper, we are using the ensemble learning technique of ML with stacking to predict the glucose(mg/dL) level of diabetic patients by using a urine dataset. For the experimental evaluation we utilized Support Vector Regressor (SVR), Decision Tree Regressor (DT), KNeighbors Regressor (KN) models, and evaluation metrics like Root Mean Square Error (RMSE), Mean Square Error (MSE), Mean Average Precision (MAP) is used to evaluate the outcomes. Throughout the assessment, it is assessed that the Ensemble model predicts the glucose level (mg/dL) with the lowest 0.09 RMSE, 0.22 MSE, and 3.1 MAP.

Keywords: Machine Learning, Patients, Urine Dataset, Diabetic patients