Design Of Neural Networks For Classification Of Electrocardiographic Signals
Giovanni Bortolan, Rosanna Degani, Jos L. Willems · 2005
The aim of this paper is to test the performance of neural networks in the classification of the conventional electrocardiogram (ECG). A large validated database of ECG signals whose classification is based on ECG-independent data has been used. The performance of the neural networks is compared with classical statistical classification techniques. Although the accuracy obtained with the neural networks is a little bit lower than in the statistical approach, the results justify further studies on the structure of the networks.