Full Program »
Ensemble-Based Blood Glucose Prediction In Diabetes Mellitus: A Review
This paper aims to conduct a systematic mapping study to review the state of the art in predicting blood glucose using ensemble methods with regard to 4 facets: years and publication sources, incorporated factors to predict blood glucose, data sources employed to collect the data, types of considered ensembles. A total of 32 papers published between 2000 and 2020 in six digital libraries were selected and reviewed with regard to the four identified mapping questions. This review results found that research in in this topic is still at its beginning but is gaining a growing interest in the last years. Past blood glucose, insulin bolus, diet and physical activity are the most frequently incorporated factors used to predict blood glucose with a prediction horizon between 5 minutes and 120 minutes. Furthermore, both homogeneous and heterogeneous ensembles have been investigated.