Research on Non-Invasive Blood Glucose Detection Method Based on PSO-GRNN

Applying the PSO (PSO) method, a GRNN-based hybrid model was established to reduce the discomfort caused by invasive blood glucose and improve its monitoring effect. forecast model. The results show that the PSO GRNN model has better forecasting performance, and Clark’s error grid analysis reveals that the model’s prediction results fall within the A region at 100% and that the root mean square error is 0.26, both of which are acceptable by clinical standards. The model is capable of swiftly and rather accurately measuring blood glucose levels.