Heart Failure diagnosis based on deep learning techniques

The aim of the study is to address the heart failure (HF) diagnosis with the application of deep learning approaches. Seven deep learning architectures are implemented, where stacked Restricted Boltzman Machines (RBMs) and stacked Autoencoders (AEs) are used to pre-train Deep Belief Networks (DBN) a...

Full description

Saved in:
Bibliographic Details
Published in:2021 43rd Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC) Vol. 2021; pp. 1757 - 1760
Main Authors: Papadopoulos, Theofilos G., Plati, Daphni, Tripoliti, Evanthia E., Goletsis, Yorgos, Naka, Katerina K., Rammos, Aidonis, Bechlioulis, Aris, Watson, Chris, McDonald, Kenneth, Ledwidge, Mark, Pharithi, Rebabonye, Gallagher, Joseph, Fotiadis, Dimitrios I.
Format: Conference Proceeding Journal Article
Language:English
Published: United States IEEE 01.11.2021
Subjects:
Algorithms
Biological system modeling
Deep Learning
Electrocardiography
Heart
Heart Failure - diagnosis
History
Hospitals
Humans
Neural Networks, Computer
Sensitivity
ISSN:2694-0604, 2694-0604
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:The aim of the study is to address the heart failure (HF) diagnosis with the application of deep learning approaches. Seven deep learning architectures are implemented, where stacked Restricted Boltzman Machines (RBMs) and stacked Autoencoders (AEs) are used to pre-train Deep Belief Networks (DBN) and Deep Neural Networks (DNN). The data is provided by the University College Dublin and the 2nd Department of Cardiology from the University Hospital of Ioannina. The features recorded are grouped into: general demographic information, physical examination, classical cardiovascular risk factors, personal history of cardiovascular disease, symptoms, medications, echocardiographic features, laboratory findings, lifestyle/habits and other diseases. The total number of subjects utilized is 422. The deep learning methods provide quite high results with the Autoencoder plus DNN approach to demonstrate accuracy 91.71%, sensitivity 90.74%, specificity 92.31% and f-score 89.36%.
Bibliography:ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 23
ISSN:2694-0604
2694-0604
DOI:10.1109/EMBC46164.2021.9630409