A Study of Extended Sample Entropy Based on the Correlation Integral

The computation between the sample entropy and correlation integral is compared. Their quantitative relationship is analyzed, and a new understanding way of sample entropy is presented. The sample entropy algorithm is extended by introducing the delay time variable of correlation integral to it and changing the variable value, by which a new way for signal feature extraction is provided. The extended sample entropy with delay time interval is analyzed with single English phoneme signals as experimental objects. The experimental results show that the extended sample entropy is larger than the traditional sample entropy, and they represent new signal features.