Profile
International Journal of Computer & Software Engineering Volume 1 (2016), Article ID 1:IJCSE-109, 8 pages
https://doi.org/10.15344/2456-4451/2016/109
Original Article
Advanced Clustering Method for Neurological Assessment Using Graph Models

Herbert F. Jelinek1,2, David J. Cornforth2 and Andrei V. Kelarev1*

1Centre for Research in Complex Systems and School of Community Health, Charles Sturt University, Albury, Australia
2Applied Informatics Research Group, University of Newcastle, University Drive, Callaghan, NSW 2308, Australia
Dr. Andrei V. Kelarev, Centre for Research in Complex Systems and School of Community Health, Charles Sturt University, Albury, Australia; E-mail: andreikelarev-charlessturtuniversity@yahoo.com
25 June 2016; 16 December 2016; 19 December 2016
Jelinek HF, Cornforth DJ, Kelarev AV (2016) Advanced Clustering Method for Neurological Assessment Using Graph Models. Int J Comput Softw Eng 1: 109. doi: https://doi.org/10.15344/2456-4451/2016/109

Abstract

This work represents an advance in sophisticated methods used to detect severe cardiac autonomic neuropathy (CAN). It applies clustering based on a graph model to ECG biosignal processing results in order to optimise classification performance. Severe CAN represents a particularly significant neurological problem in diabetes healthcare as it requires urgent intervention to reduce the risk of sudden cardiac death. The introduction of a new Clustering System Based on Graphs (CSBG) combined with heart rate features determined from recorded ECG biosignals was intended as a means of enhancing the effectiveness of the diagnosis of severe CAN. Here we study a novel heart rate descriptor – Allan exponents (AE) to determine the effectiveness of CSBG and compare the results with performance of other classification and clustering systems available in Sage. The best outcomes were obtained by CSBG in combination with AE, which improved the F measure of classification performance to 0.92 and outperformed several other classification and clustering systems in our experiments.