Journal
Scientific and technical journal «Priborostroenie»
UDK004.93
Issue:12 (67)
Download PDF587 Kbyte
The relevance of the problem of selecting informative features that allow detecting possible cardiac arrhythmias and classifying arrhythmias is shown. It is proposed to use eight informative features of the electrocardiocomplex. It is shown that these features can be reduced to two aggregated indicators. The study was conducted on the materials of lead II of the open verified MIT-BIH Arrhythmia DB database and the database of multichannel ECGs of the St. Petersburg Institute of Cardiology Technology “INKART”. The results of the classification using the support vector method demonstrate the validity of the selected space of informative features for classifying various states of the cardiac rhythm.