Martin Lagerholm, Carsten Peterson, Guido Braccini, Lars Edenbrandt and Leif Sörnmo
Clustering ECG complexes using Hermite functions and self-organizing maps
IEEE Transactions on Biomedical Engineering 47, 838-848 (2000)
An integrated method for identifying and classifying ECG complexes in the MIT-BIH Arrhythmia Database is presented. The QRS complexes are extracted and the RR intervals are calculated in an autonomous way. Each complex is subsequently decomposed into Hermite basis functions, from which the resulting coefficients and widths are used to represent the beats. By means of this representation, unsupervised self-organizing neural networks are employed to cluster the data into 25 groups. The resulting clusters exhibit a very low degree of misclassification (1.5%). The approach outperforms published supervised learning methods using the same data. The method is also successfully compared with a more conventional template cross-correlation clustering method.