Robust multiple cardiac arrhythmia detection through bispectrum analysis

Authors:
A. Lanatá;G. Valenza;C. Mancuso;E. P. Scilingo
Affiliations:
Department of Information Engineering and Interdepartmental Research Center "E. Piaggio", University of Pisa, Via Caruso 16, 56122 Pisa, Italy;Department of Information Engineering and Interdepartmental Research Center "E. Piaggio", University of Pisa, Via Caruso 16, 56122 Pisa, Italy;Department of Information Engineering and Interdepartmental Research Center "E. Piaggio", University of Pisa, Via Caruso 16, 56122 Pisa, Italy;Department of Information Engineering and Interdepartmental Research Center "E. Piaggio", University of Pisa, Via Caruso 16, 56122 Pisa, Italy
Venue:
Expert Systems with Applications: An International Journal
Year:
2011

Citing 8
Cited 1

Glossary of Terms

Machine Learning - Special issue on applications of machine learning and the knowledge discovery process
Statistical Pattern Recognition: A Review

IEEE Transactions on Pattern Analysis and Machine Intelligence
Pattern Classification (2nd Edition)

Pattern Classification (2nd Edition)
Frequency-domain features for ECG beat discrimination using grey relational analysis-based classifier

Computers & Mathematics with Applications
Adaptive wavelet network for multiple cardiac arrhythmias recognition

Expert Systems with Applications: An International Journal
On convergence properties of the em algorithm for gaussian mixtures

Neural Computation
Pattern Recognition Using Invariants Defined From Higher Order Spectra- One Dimensional Inputs

IEEE Transactions on Signal Processing
Nearest neighbor pattern classification

IEEE Transactions on Information Theory

ECG arrhythmia classification based on optimum-path forest

Expert Systems with Applications: An International Journal

Quantified Score

Hi-index	12.05

Visualization

Abstract

This paper investigates the use of Higher Order Spectra parameters to identify the most common multiple cardiac arrhythmias. In detail, we calculated magnitude of bispectrum, three values of bispectrum entropy, mean and variance of the phase of bispectrum integrated over a particular region wherein no bispectrum aliasing is assumed. This set of features is used to distinguish normal QRS from five different classes of arrhythmia over a large amount of normal and pathologic ECG signals. An accurate parametric and non-parametric analysis of these feature distributions is also performed in order to identify the optimum classifier. Experimental tests were performed on signals gathered from the MIT-BIH Arrhythmias Database, and mean and standard deviation of all confusion matrixes obtained from 20 steps of cross validation are shown. Results showed that the bispectrum is high performance, reliable and robust method to identify arrhythmias.