Title :
Classification of ECG waveforms by using RCE neural network and genetic algorithms
Author_Institution :
Fac. of Electr. & Electron., Istanbul Tech. Univ., Turkey
fDate :
8/28/1997 12:00:00 AM
Abstract :
A modified restricted Coulomb energy (RCE) neural network is proposed for electrocardiogram (ECG) waveform classification. In the new algorithm, the locations of the centres of the hyperspheres are optimised in order to minimise the number of hyperspheres used and to improve classification performance. Optimisation is carried out with the genetic algorithm
Keywords :
electrocardiography; genetic algorithms; medical signal processing; neural nets; pattern classification; waveform analysis; electrocardiogram waveform classification; genetic algorithm; hypersphere; optimisation; restricted Coulomb energy neural network;
Journal_Title :
Electronics Letters
DOI :
10.1049/el:19971019
