Wavelet transform and vector machines as emerging tools for computational medicine

Electrocardiogram (ECG) is a most primitive and important test to analyse the status of the heart functioning. During this test, different types of noises and artefacts get involved in the captured electrical signal which affects the performance of overall diagnosis. In general, computer aided decision system (CADS) performs three operations viz. pre-processing, feature extraction and classification to reach a decision for analyzing an ECG signal. Among three waves of an ECG signal, QRS-complex is to be examined most critically to diagnose existence of a possible cardiovascular disease. But detection of exact locations of QRS complexes is still a challenging task as they are hidden by various noises and artefacts. Therefore, in this paper emerging tools such as wavelet transform (WT), adaptive autoregressive modelling (AARM) and vector machines (VMs) like support vector machine (SVM) and relevance vector machine (RVM) are proposed to be used for pre-processing, feature extraction and classification, respectively for utilizing distinct advantages of each. For instance, WT provides better time–frequency resolution, AARM possesses parameters that vary with time leading to the measurement of time-varying spectra and VMs models the non-linear data stably. Also, RVM has been proposed to be used for the first time here for ECG signal analysis as it needs much less kernel functions. SVM has been used for comparison purpose only. The performance of the proposed methodology is evaluated on the basis of widely used performance parameters such as sensitivity (Se), positive predictivity (Pp), accuracy (Acc) and detection rate (Dr). Highlight of the proposed methodology is that it yields consistently high values of all the widely used and critical performance parameters i.e. Se of 99.95%, Pp of 99.95%, Dr of 99.95%, and Acc of 99.93%. These results are highest amongst other techniques existing in the literature, indicating its usefulness for handling real-time heart related emergent cases. © 2023, The Author(s), under exclusive licence to Springer-Verlag GmbH Germany, part of Springer Nature.