Hierarchical Clustering Method for Bayès Syndrome Detection

Authors

  • Franco, Lorena Gisela
  • Escobar, Luis A.
  • Wainschenker, Rubén
  • Bayès de Luna, Antoni
  • Massa, José M.

DOI:

https://doi.org/10.18687/LACCEI2023.1.1.1314

Keywords:

Síndrome de Bayès, ECG, Árbol Jerárquico, K Means++, FAUM

Abstract

Bayès Syndrome manifests itself in the cardiac cycle of an electrocardiogram. It presents associations with multiple medical conditions, and the detection at an early stage is of particular interest. In this article, the Hierarchical Clustering method was applied with the implementation of Matlab to identify each signal in 4 groups or categories of interest for diagnosing Bayès Syndrome. Different configuration values were explored for the linkage parameter. The best result was obtained with the 'ward' option with a normalized sample signal in amplitude and time, achieving a total f1 Score of 0.88. The performance of Hierarchical Clustering was compared to the one reached with K-Means++ and FAUM methods from previous works for signals normalized in amplitude. The Hierarchical Clustering total f1 Score indicator was lower than the value obtained from the two K Means++ implementations and higher than the adjusted FAUM value.

Downloads

Published

2024-04-16

Issue

Section

Articles