Classification of Cardiac Arrhythmias by Convolutional Neural Networks and Particle Swarm Optimization

A cardiac arrhythmia is an irregular heartbeat that results in an abnormal electrical impulse, and its type is defined by its rhythm and duration. Its classification has been addressed in different fields of science, highlighting the use of deep learning algorithms (DLA). This research used a hybrid...

Descripción completa

Detalles Bibliográficos
Autores: Santander-Baños, Fredy, Hernández-Romero, Norberto, Barragán-Vite, Irving, Karelin, Oleksandr, Medina-Marín, Joselito
Tipo de recurso: artículo
Estado:Versión publicada
Fecha de publicación:2022
País:México
Institución:UNIVERSIDAD AUTÓNOMA DEL ESTADO DE HIDALGO
Repositorio:PÄDI Boletín Científico de Ciencias Básicas e Ingeniería del ICBI
Idioma:español
OAI Identifier:oai:repository.uaeh.edu.mx:article/8655
Acceso en línea:https://repository.uaeh.edu.mx/revistas/index.php/icbi/article/view/8655
Access Level:acceso abierto
Palabra clave:Convolutional Neural Networks
Particle Swarm Optimization
Computational Model
Cardiac Arrhythmias Classification
Electrocardiograms
Redes Neuronales Convolucionales
Optimización por Enjambre de Partículas
Modelo Computacional
Clasificación de Arrimitas Cardíacas
Electrocardiogramas
Descripción
Sumario:A cardiac arrhythmia is an irregular heartbeat that results in an abnormal electrical impulse, and its type is defined by its rhythm and duration. Its classification has been addressed in different fields of science, highlighting the use of deep learning algorithms (DLA). This research used a hybrid model between Convolutional Neural Networks (CNN) and the Particle Swarm Optimization (PSO) metaheuristic algorithm; for the classification of cardiac arrhythmias. The metaheuristic was in charge of optimizing the architecture of the neural network layers, through the minimization of the loss during training and testing. Data were obtained from the MIT-BIH Arrhythmia dataset, which describes five categories of arrhythmias. The results achieved showed that the metaheuristic is a reliable algorithm in the search for the best layer architecture, achieving an accuracy of 97%, which means that the use of metaheuristic techniques is an option that should be taken into consideration when to optimize the performance of convolutional neural networks.