Abstract

Artificial neural networks have been successfully applied to a wide range of problems. In pat-
tern recognition, they have been used in several tasks, such as feature extraction, dimension
reduction, and classification. In this work, we propose two ECG heartbeat classification models
based on collaborating different types of neural networks. The main aim is to combine their
Complementary properties.
The first model uses a stacked sparse autoencoder(SSAE)as feature extractor and a system
of multipleMulti-layeredperceptrons(MLP)as a classifier. In this model, the entire problem
is divided into simpler parts, which are resolved using different MLPs. The second model also
uses a SSAE to extract features in addition to two other dynamic features. In this model, the
classification is performed by a hybrid neural model based on combining random and RBF neural
networks.
The proposed models are evaluated on the MIT-BIH arrhythmia dataset. The tests are based on
the inter-patient paradigm, in which the training and test data are taken from different patients.
The obtained results are compared with some of the state-of-the-art methods.
Keywords: Pattern recognition, classification, neural networks, machine learning, deep learn-
ing, ECG dataset.

Information

Item Type:	Thesis
Divisions:	» Faculty of Science and Technology
ePrint ID:	4431
Date Deposited:	2023-09-17
Further Information:	Google Scholar
URI:	https://www.univ-soukahras.dz/en/publication/article/4431

BibTex