PREDIABETES TEST SYSTEM AND METHOD BASED ON COMBINED HEART AND BRAIN ELECTRICAL INFORMATION

摘要

A prediabetes test system and method based on combined heart and brain electrical information, the system comprising: a signal acquisition module (110), which utilizes a wearable device to noninvasively and synchronously obtain a heart electrical signal and a brain electrical signal; a feature extraction module (120), which utilizes multiple means to perform dimensionality reduction on a combined feature set made up of heart electrical features and brain electrical features, obtaining a plurality of dimensionally reduced combined feature sets, then selecting heart electrical features and brain electrical features that satisfy a preset correlation standard by means of analyzing correlations of the plurality of dimensionally reduced combined feature sets and a blood glucose concentration value, and forming an optimized combined feature set; and a multi-model merging module (130), used for respectively inputting the optimized combined feature set into multiple types of neural network models that have undergone training, and obtaining a prediabetes test result for a user by means of merging output results from the multiple types of neural networks. The present method and system can provide a noninvasive, painless, quick, convenient, comfortable, and low-cost prediabetes test solution.