A Prediction Method for Blood Glucose Based on Grey Wolf optimization Evolving Kernel Extreme Learning Machine

Abstract

Patients with type 1 diabetes need to acquire their blood glucose prediction values which can make sure the actual value is well controlled within the normal range. Therefore, the accuracy of the blood glucose prediction method is very important. In this paper, the radial basis function kernel extreme learning machine (KELM) is used to predict blood glucose, and the parameters are adjusted by the grey wolf optimization (GWO) algorithm. Experiment results show that KELM based on GWO algorithm has great robustness and better generalization performance comparison with traditional extreme learning machine algorithm. The GWO-KELM model achieved high prediction accuracy.