Blood Glucose Prediction Based on Empirical Mode Decomposition and SSA-KELM

Predicting the blood glucose of type 1 diabetes patients in advance is of great significance to the prevention and treatment of diabetes. The prediction model used in this paper is based on the Sparrow Search Algorithm (SSA) of Empirical Mode Decomposition (EMD) to optimize the Kernel Extreme Learning Machine (KELM). In this work, EMD is used to decompose blood glucose values into different frequency sequences. Secondly, SSA- KELM is trained and each sub-sequence is predicted separately, and finally the prediction sequence is reconstructed to obtain the predicted value of blood glucose. The experimental results show that the SSA-KELM model has higher prediction accuracy than the ELM and PSO-KELM algorithms, and can be used for blood glucose prediction models.