Classification of diabetic neuropathic patients from emg data: a machine learning approach

It is of interest for health professionals to automatically classify and diagnose diabetic neuropathy (DPN) patients. Therefore, in this study, we initially propose a machine learning based tool capable of discriminating healthy from DPN individuals using motor data from surface electromyography signals. The tool is based on deep learning techniques, a convolutional neural network applied to fuzzy entropy extracted from high-density surface electromyography (sEMG). The tool showed ~75% of accuracy with cross validation, revealing a promising method for DPN diagnose in a clinical context, even if the professional does not have experience.