Online Neural Network Training for Automatic Ischemia Episode Detection

Abstract

Abstract. Myocardial ischemia is caused by a lack of oxygen and nutrients to the contractile cells and may lead to myocardial infarction with its severe consequence of heart failure and arrhythmia. An electrocardiogram (ECG) represents a recording of changes occurring in the electrical potentials between different sites on the skin as a result of the cardiac activity. Since the ECG is recorded easily and non-invasively, it becomes very important to provide means of reliable ischemia detection. Ischemic changes of the ECG frequently affect the entire repolarization wave shape. In this paper we propose a new classification methodology that draws from the disciplines of clustering and artificial neural networks, and apply it to the problem of myocardial ischemia detection. The results obtained are promising.

Novel Approaches to Unsupervised Clustering Through k-Windows Algorithm

“…With the incorporation of computational geometry techniques the algorithm achieves a comparatively low time complexity. The algorithm has been successfully applied in numerous applications including bioinformatics [47,48], medical diagnosis [31,49], time series prediction [35] and web personalization [41].…”

Unsupervised Clustering Using Fractal Dimension

“…On the other hand, for low dimensional data with a large number of points the approach of Alevizos [1998] appears more attractive. The unsupervised k-windows algorithm has been successfully applied in numerous applications including bioinformatics [Tasoulis et al, 2004a[Tasoulis et al, , 2004b, medical diagnosis [Magoulas et al, 2004;Tasoulis et al, 2003], time series prediction [Pavlidis et al, 2003] and web personalization [Rigou et al, 2004].…”

Density Based Text Clustering

“…Furthermore, the algorithm can detect clusters of arbitrary shapes and determine the cluster number without additional computational burden. Although it has already been applied in numerous tasks that range from medical diagnosis [9,14], to web personalization [12], it has not been used yet on text data. In this paper, through indicative results of the UkW algorithm, we demonstrate that the feature scoring metrics have to be suitably exploited, so that the cluster detection ability of UkW is optimized.…”