A Review on Time Series Motif Discovery Techniques an Application to ECG Signal Classification: ECG Signal Classification Using Time Series Motif Discovery Techniques

A Review on Time Series Motif Discovery Techniques an Application to ECG Signal Classification: ECG Signal Classification Using Time Series Motif Discovery Techniques

Ramanujam Elangovan (Department of Information Technology, Thiagarajar College of Engineering, Madurai, Tamil Nadu, India) and Padmavathi S. (Department of Information Technology, Thiagarajar College of Engineering, Madurai, Tamil Nadu, India)
DOI: 10.4018/IJAIML.2019070103

Abstract

Cardiovascular disease diagnosis from an ECG signal plays an important and significant role in the health care system. Recently, numerous researchers have developed an automatic time series-based multi-step diagnosis system for the fast and accurate diagnosis of ECG abnormalities. The multi-step procedure involves ECG signal acquisition, signal pre-processing, feature extraction, and classification. Among which, the feature extraction plays a vital role in the field of accurate diagnosis. The features may be different types such as statistical, morphological, wavelet or any other signal-based approach. This article discusses various time series motif-based feature extraction techniques with respect to a different dimension of ECG signal.
Article Preview
Top

Ecg Signal Acquisition

ECG recordings of different subjects can be acquired from different publicly available databases (Goldberger et al, 2000 and chen et al, 2015) containing various pathological and normal conditions. Signals of these databases are available freely and they permit the creation of standardization for the evaluation of automatic cardiac conditions. The standardization has been developed by AAMI standards (Association for the Advancement of Medical Instrumentation) in ANSI/ AAMI EC 57:1998/2008. The list of publicly available databases recommended by the standardization is available in Physionet (Goldberger et al, 2000) as shown in Table 1. One of the familiar databases used with time series motif discovery problem is MIT-BIH Arrhythmia database. The detailed descriptions of those data are available in Goldberger et al, 2000. Further, most of the time series motif discovery technique use UCR archive (chen et al, 2015). The UCR archive contains more than 100 time series datasets and the detailed description of data related to ECG signals are shown in Table 2.

Complete Article List

Search this Journal:
Reset
Open Access Articles: Forthcoming
Volume 10: 2 Issues (2020): 1 Released, 1 Forthcoming
Volume 9: 2 Issues (2019)
View Complete Journal Contents Listing