Sleep Stage Estimation Using ECG

Sleep stage judgement is not easy because it is difficult to measure biological signals. In addition, most people don’t know how to judge sleep stage. Therefore, this paper describes a method for estimation sleep stage using only an ECG. The used parameters from ECG was Poincare plot and DFA. The estimation model was decision tree. As a result, the best score achieved 74% accuracy. The accuracy of all other subjects achieved above 60%. The precision of WAKE and LIGHT all exceeded 50%. The estimation of REM was not working, but we are considering the accuracy is further improved by using the processing of individual differences and frequency analysis.