A Robust Martingale Approach for Detecting Abnormalities in Human Heartbeat Rhythm

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review


The analysis of electrocardiogram data is vital to the healthcare system to improve and monitor health conditions. Existing algorithms are effective in discovering abnormalities in electrocardiogram data streams but most of these approaches do not focus on the intensity and duration of these anomalies. In this paper, we propose a new method called alignment of the martingale sequence (AMS) that improves previous approaches using dynamic time warping and particle swarm optimisation to obtain the optimal parameter that maximises F1. Our proposed method can also estimate the severity and extent of an abnormal heartbeat rate. Experimental results show that the proposed approach makes some improvements over the traditional method.

This paper won the Best Paper Award
Original languageEnglish
Title of host publication7th Collaborative European Research Conference
Subtitle of host publicationCERC2021
PublisherCEUR Workshop Proceedings
Publication statusAccepted/In press - 23 Aug 2021
Event7th Collaborative European Research Conference (CERC 2021) - Cork, Cork, Ireland
Duration: 9 Sept 202110 Sept 2021


Conference7th Collaborative European Research Conference (CERC 2021)
Internet address


  • heart rate
  • Dynamic time warping
  • ECG sequence
  • martingales


Dive into the research topics of 'A Robust Martingale Approach for Detecting Abnormalities in Human Heartbeat Rhythm'. Together they form a unique fingerprint.

Cite this