Yucelbas S.Tezel G.2020-03-262020-03-2620139.78973E+12https://hdl.handle.net/20.500.12395/30204IADIS International Conference Intelligent Systems and Agents 2013, ISA 2013, IADIS European Conference on Data Mining 2013, ECDM 2013, Part of the IADIS Multi Conference on Computer Science and Information Systems 2013, MCCSIS 2013 -- 22 July 2013 through 24 July 2013 -- Prague -- 100179Artificial neural networks, used for practices of engineering, are strong tool for useful relationships between data. Particle swarm optimization is successfully carried out to train feedforward neural networks. In this study, heart diseases diagnosis was realized by using Electrocardiography (ECG) records that taken from MIT-BIH ECG database, which were used to classify 6 different arrhythmias for training and testing. These are normal sinus rhythm (NSR), ventricular tachycardia (VTK), sinus arrhythmia (SA), atrial premature contraction (APC), atrial fibrillation (AF) and ventricular trigeminy (VTI). Artificial Neural Networks (ANN) and Artificial Neural Networks trained by particle swarm optimization (PSO-NN) are used as a classifier and they are compared. Experimental results have revealed that PSO-NN model is better for classification of ECG signals than traditional ANN. © 2013 IADIS.eninfo:eu-repo/semantics/closedAccessArtificial neural networksECGHeart disease diagnosisParticle swarm optimizationConference Object5966N/A