Detection of Cardiovascular Diseases Using Predictive Models Based on Deep Learning Techniques: A Hybrid Neutrosophic AHP-TOPSIS Approach for Model Selection

In Ecuador and globally, cardiovascular diseases are the leading cause of mortality, accounting for a worrying 26.49% of deaths in 2019. An approach based on deep learning is applied to improve the capacity for early prediction and reduce its incidence. In this work, three different models were prop...

Full description

Saved in:
Bibliographic Details
Main Authors: Julio Barzola-Monteses, Rosangela Caicedo-Quiroz, Franklin Parrales-Bravo, Cristhian Medina-Suarez, Wendy Yanez-Pazmino, David Zabala-Blanco, Maikel Y. Leyva-Vazquez
Format: Article
Language:English
Published: University of New Mexico 2024-12-01
Series:Neutrosophic Sets and Systems
Subjects:
Online Access:https://fs.unm.edu/NSS/17CardiovascularDiseases.pdf
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849228937464381440
author Julio Barzola-Monteses
Rosangela Caicedo-Quiroz
Franklin Parrales-Bravo
Cristhian Medina-Suarez
Wendy Yanez-Pazmino
David Zabala-Blanco
Maikel Y. Leyva-Vazquez
author_facet Julio Barzola-Monteses
Rosangela Caicedo-Quiroz
Franklin Parrales-Bravo
Cristhian Medina-Suarez
Wendy Yanez-Pazmino
David Zabala-Blanco
Maikel Y. Leyva-Vazquez
author_sort Julio Barzola-Monteses
collection DOAJ
description In Ecuador and globally, cardiovascular diseases are the leading cause of mortality, accounting for a worrying 26.49% of deaths in 2019. An approach based on deep learning is applied to improve the capacity for early prediction and reduce its incidence. In this work, three different models were proposed and compared: deep neural networks (DNN), convolutional neural networks (CNN), and multilayer perceptron (MLP). Experiments were conducted in two scenarios: one using a dataset that included 12 variables, and another in which the variables were reduced to those most significantly correlated with cardiovascular disease, i.e., 4 variables; both scenarios with 918 clinical records per variable. Using the Neutrosophic AHP-TOPSIS method for model selection, the CNN model trained with the original dataset was identified as the best-performing model among the proposed options. In specific terms, the evaluation metrics of the CNN model were as follows: an accuracy of 92.17%, a sensitivity of 94.51%, a specificity of 90.78%, an F1-Score of 93.30%, and an area under the ROC curve of 90.03%.
format Article
id doaj-art-67cba5e6805f41de82592807cb7930f7
institution Kabale University
issn 2331-6055
2331-608X
language English
publishDate 2024-12-01
publisher University of New Mexico
record_format Article
series Neutrosophic Sets and Systems
spelling doaj-art-67cba5e6805f41de82592807cb7930f72025-08-22T11:00:41ZengUniversity of New MexicoNeutrosophic Sets and Systems2331-60552331-608X2024-12-017421022610.5281/zenodo.14418335Detection of Cardiovascular Diseases Using Predictive Models Based on Deep Learning Techniques: A Hybrid Neutrosophic AHP-TOPSIS Approach for Model SelectionJulio Barzola-MontesesRosangela Caicedo-QuirozFranklin Parrales-BravoCristhian Medina-SuarezWendy Yanez-PazminoDavid Zabala-BlancoMaikel Y. Leyva-VazquezIn Ecuador and globally, cardiovascular diseases are the leading cause of mortality, accounting for a worrying 26.49% of deaths in 2019. An approach based on deep learning is applied to improve the capacity for early prediction and reduce its incidence. In this work, three different models were proposed and compared: deep neural networks (DNN), convolutional neural networks (CNN), and multilayer perceptron (MLP). Experiments were conducted in two scenarios: one using a dataset that included 12 variables, and another in which the variables were reduced to those most significantly correlated with cardiovascular disease, i.e., 4 variables; both scenarios with 918 clinical records per variable. Using the Neutrosophic AHP-TOPSIS method for model selection, the CNN model trained with the original dataset was identified as the best-performing model among the proposed options. In specific terms, the evaluation metrics of the CNN model were as follows: an accuracy of 92.17%, a sensitivity of 94.51%, a specificity of 90.78%, an F1-Score of 93.30%, and an area under the ROC curve of 90.03%.https://fs.unm.edu/NSS/17CardiovascularDiseases.pdfheart diseasepredictionconvolutional neural networkdeep neural networkmultilayer perceptronneutrosophic ahp-topsis
spellingShingle Julio Barzola-Monteses
Rosangela Caicedo-Quiroz
Franklin Parrales-Bravo
Cristhian Medina-Suarez
Wendy Yanez-Pazmino
David Zabala-Blanco
Maikel Y. Leyva-Vazquez
Detection of Cardiovascular Diseases Using Predictive Models Based on Deep Learning Techniques: A Hybrid Neutrosophic AHP-TOPSIS Approach for Model Selection
Neutrosophic Sets and Systems
heart disease
prediction
convolutional neural network
deep neural network
multilayer perceptron
neutrosophic ahp-topsis
title Detection of Cardiovascular Diseases Using Predictive Models Based on Deep Learning Techniques: A Hybrid Neutrosophic AHP-TOPSIS Approach for Model Selection
title_full Detection of Cardiovascular Diseases Using Predictive Models Based on Deep Learning Techniques: A Hybrid Neutrosophic AHP-TOPSIS Approach for Model Selection
title_fullStr Detection of Cardiovascular Diseases Using Predictive Models Based on Deep Learning Techniques: A Hybrid Neutrosophic AHP-TOPSIS Approach for Model Selection
title_full_unstemmed Detection of Cardiovascular Diseases Using Predictive Models Based on Deep Learning Techniques: A Hybrid Neutrosophic AHP-TOPSIS Approach for Model Selection
title_short Detection of Cardiovascular Diseases Using Predictive Models Based on Deep Learning Techniques: A Hybrid Neutrosophic AHP-TOPSIS Approach for Model Selection
title_sort detection of cardiovascular diseases using predictive models based on deep learning techniques a hybrid neutrosophic ahp topsis approach for model selection
topic heart disease
prediction
convolutional neural network
deep neural network
multilayer perceptron
neutrosophic ahp-topsis
url https://fs.unm.edu/NSS/17CardiovascularDiseases.pdf
work_keys_str_mv AT juliobarzolamonteses detectionofcardiovasculardiseasesusingpredictivemodelsbasedondeeplearningtechniquesahybridneutrosophicahptopsisapproachformodelselection
AT rosangelacaicedoquiroz detectionofcardiovasculardiseasesusingpredictivemodelsbasedondeeplearningtechniquesahybridneutrosophicahptopsisapproachformodelselection
AT franklinparralesbravo detectionofcardiovasculardiseasesusingpredictivemodelsbasedondeeplearningtechniquesahybridneutrosophicahptopsisapproachformodelselection
AT cristhianmedinasuarez detectionofcardiovasculardiseasesusingpredictivemodelsbasedondeeplearningtechniquesahybridneutrosophicahptopsisapproachformodelselection
AT wendyyanezpazmino detectionofcardiovasculardiseasesusingpredictivemodelsbasedondeeplearningtechniquesahybridneutrosophicahptopsisapproachformodelselection
AT davidzabalablanco detectionofcardiovasculardiseasesusingpredictivemodelsbasedondeeplearningtechniquesahybridneutrosophicahptopsisapproachformodelselection
AT maikelyleyvavazquez detectionofcardiovasculardiseasesusingpredictivemodelsbasedondeeplearningtechniquesahybridneutrosophicahptopsisapproachformodelselection