Heart disease detection using ECG lead I and multiple pattern recognition classifiers

Renato Pereira, Bruno Bispo, Pedro Miguel Rodrigues

Research output: Contribution to journalArticlepeer-review

2 Downloads

Abstract

ECG is an important tool to assist in heart diseases diagnosis. The works found in the literature have the common goal of discriminating between binary study groups, one pathological and one control, even when ECG records from patients diagnosed with several pathologies are available in the databases. This work proposes a method to detect ECG morphological features and to analyze the capacity of this ECG features to discriminate 28 pairs of study groups, combining 7 pathological groups and 1 control group, presented in the PTB Diagnostic ECG Database. For each pair, it was achieved an accuracy between 77.4% and 100%, with an average of 94%, using several pattern recognition classifiers.
Original languageEnglish
Pages (from-to)1-8
Number of pages8
JournalIOSR Journal of Engineering
Volume10
Issue number4
Publication statusPublished - 4 Apr 2020

Keywords

  • Heart diseases
  • ECG features
  • Pattern recognition
  • PTB Diagnostic ECG databases
  • Classifiers

Fingerprint

Dive into the research topics of 'Heart disease detection using ECG lead I and multiple pattern recognition classifiers'. Together they form a unique fingerprint.

Cite this