An effective fault detection method with FDA classifier and global model for continuous glucose monitor (CGM)

Abstract

Continuous glucose monitoring (CGM) plays an important role in the treatment of diabetes. False information of glucose, caused by the faults of CGM, will reduce the treatment effect, and even endanger the lives of the diabetic patients. In order to avoid the impact of the glucose fluctuations, a classification-based method is used to detect CGM faults. However, the poor classification algorithm in this method reduces the performance of fault detection. In this study, we will analyze the classified glucose data and improve the fault-detect method by building the FDA classifier. On the basis of the classifier, the results of PCA models are fused for the final fault detection. Besides, the global fault-detect model is studied in the paper to reducing the modeling cost. The simulation results show that the improved fault-detect method we presented has a better performance in CGM fault detection.