SVM Based Chronic Fatigue Syndrome Evaluation for Intelligent Garment

Abstract

Chronic fatigue syndrome (CFS) also called sub-health is a serious and complex problem for modern people all over the world. But the methods of CFS diagnosis up to now are very elementary. This paper tries to establish a CFS evaluation model based on human body's vital signals, especially ECG. Firstly, an intelligent garment oriented physiological signal capturing and processing platform is proposed. Then, a multi-class SVM-based strategy to render a diagnosis between various degrees of CFS is constructed. Based on the ISNI-DHU CFS database we set up, the results show that the diagnosis model achieve high classification accuracy, at 97.4% of average accuracy, and heartbeat parameters can be effectively used to evaluation of CFS.