Electrocardiogram classification using delay differential equations.

摘要: Time series analysis with nonlinear delay differential equations (DDEs) reveals as well spectral properties of the underlying dynamical system. Here, global DDE models were used to analyze 5 min data segments electrocardiographic (ECG) recordings in order capture distinguishing features for different heart conditions such normal beat, congestive failure, and atrial fibrillation. The number terms delays model nonlinearity have be selected that are most discriminative. form best separates three classes was chosen by exhaustive search up third polynomials. Such an approach can provide deep insight into nature since linear a correspond main time-scales signal related couplings between harmonic parts. DDEs able detect fibrillation accuracy 72%, failure 88%, beat 97% from 5 min ECG, much shorter time interval than required achieve comparable performance other methods.