Cardiovascular diseases (CVDs) are a class of medical conditions characterized by impaired functionality of the heart and blood vessels, among which coronary heart disease, cerebrovascular diseases, and heart failure are prominent. These diseases account for approximately one-third of all deaths recorded worldwide. A significant challenge currently faced by clinicians is the early prediction of heart failure. The employment of Machine Learning techniques has emerged as a leading method in predicting patient survival who exhibit symptoms of heart failure. Nonetheless, achieving a significant level of accuracy remains a topic that demands further investigation. The present study introduces a novel approach that leverages Reinforcement Learning (RL) to enhance the performance of the artificial neural network (ANN) techniques for survival prediction by identifying the optimal configuration of model hyper-parameters (HPO). The proposed approach has been assessed using a conventional benchmark dataset comprising the medical records of 299 patients. The results were compared with recent ANN approaches working on the same dataset. The findings indicate that the proposed approaches enhance the performance of the ANN-based heart failure predictive model.

Hyper-Parameter Optimization through Reinforcement Learning for Survival Prediction of Patients with Heart Failure

Ribino P.
Primo
;
Di Napoli C.;Paragliola G.;Serino L.
2024

Abstract

Cardiovascular diseases (CVDs) are a class of medical conditions characterized by impaired functionality of the heart and blood vessels, among which coronary heart disease, cerebrovascular diseases, and heart failure are prominent. These diseases account for approximately one-third of all deaths recorded worldwide. A significant challenge currently faced by clinicians is the early prediction of heart failure. The employment of Machine Learning techniques has emerged as a leading method in predicting patient survival who exhibit symptoms of heart failure. Nonetheless, achieving a significant level of accuracy remains a topic that demands further investigation. The present study introduces a novel approach that leverages Reinforcement Learning (RL) to enhance the performance of the artificial neural network (ANN) techniques for survival prediction by identifying the optimal configuration of model hyper-parameters (HPO). The proposed approach has been assessed using a conventional benchmark dataset comprising the medical records of 299 patients. The results were compared with recent ANN approaches working on the same dataset. The findings indicate that the proposed approaches enhance the performance of the ANN-based heart failure predictive model.
2024
Istituto di Calcolo e Reti ad Alte Prestazioni - ICAR - Sede Secondaria Napoli
Self-Adaptive Systems,Reinforcement Learning,Survival Prediction,Heart Failure
File in questo prodotto:
File Dimensione Formato  
1-s2.0-S1877050924015989-main.pdf

non disponibili

Tipologia: Versione Editoriale (PDF)
Licenza: Altro tipo di licenza
Dimensione 1.22 MB
Formato Adobe PDF
1.22 MB Adobe PDF   Visualizza/Apri   Richiedi una copia

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/20.500.14243/519528
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus 0
  • ???jsp.display-item.citation.isi??? ND
social impact