A new application of rough set to ECG recognition

When recognizing and analyzing wave signals in electrocardiogram (ECG) automatically, present methods have some limitation and do not meet physicians' expectations. A new application of rough set to ECG recognition is presented. Firstly the recognition rules for characteristic points in ECG are reduced by using rough set theory. Then the reduced rules are used as restriction conditions of eigenvalue determination arithmetic which is presented in this paper and employed in recognizing characteristic points in ECG. There are some discussions about correlative arithmetic such as sizer method, difference method and how to choose difference parameters. Also in this paper, an example of this new application is discussed. We adopt MIT-BIH data to verify R wave recognition and the detection rate is proved higher than the one of routine recognition method.