Title: Cognitive amalgam with a distributed fuzzy case-based reasoning system for an accurate cardiac arrhythmias diagnosis

Authors: Abdeldjalil Khelassi; Mohammed Amine Chikh

Addresses: Computer Sciences Department, Abou Bakr Belkaid University, BP 230 Tlemcen, Algeria ' Biomedical Engineering Department, Abou Bakr Belkaid University, BP 230 Tlemcen, Algeria

Abstract: The distributed case-based reasoning consists of distributing the reasoning through a set of agents and the cases through a set of case bases for improving the performance and for ensuring the maintainability of the reasoning system. The most important challenge of the medical applications resides in the precision of decisions; hence the novel decision support systems should introduce more aspects of reasoning for achieving this criterion. In this paper, we describe a cognitive amalgam with a distributed case-based reasoning enriched by an expert system which infer from an XML knowledge base extracted from the domain expertise. Each cognitive agent capitalises its knowledge from same class cases by using a machine learning algorithm enriched by a fuzzy similarity measures function. The aim of this proposition is to reach the optimal criteria of medical applications. The achieved experimental results applied in the diagnosis of complex type of data obtained from the electrocardiogram (ECG) signal improves our proposition and promotes more realisations in the medical domain.

Keywords: case-based reasoning; CBR; fuzzy sets; fuzzy logic; expert systems; machine learning; distributed reasoning; electrocardiograms; ECG; cardiac arrhythmia diagnosis; decision support systems; DSS; fuzzy similarity; irregular heartbeat; abnormal heart rhythms.

DOI: 10.1504/IJICT.2015.070297

International Journal of Information and Communication Technology, 2015 Vol.7 No.4/5, pp.348 - 365

Received: 11 May 2013
Accepted: 15 Jul 2013

Published online: 02 Jul 2015 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article