Superiority of Classification Tree versus Cluster, Fuzzy and Discriminant Models in a Heartbeat Classification System.

Roger Abächerli,Miguel A Fernandez,Irena Jekova,Remo Leber,Ramun Schmid,Vessela Krasteva

doi:10.1371/journal.pone.0140123

Abstract

This study presents a 2-stage heartbeat classifier of supraventricular (SVB) and ventricular (VB) beats. Stage 1 makes computationally-efficient classification of SVB-beats, using simple correlation threshold criterion for finding close match with a predominant normal (reference) beat template. The non-matched beats are next subjected to measurement of 20 basic features, tracking the beat and reference template morphology and RR-variability for subsequent refined classification in SVB or VB-class by Stage 2. Four linear classifiers are compared: cluster, fuzzy, linear discriminant analysis (LDA) and classification tree (CT), all subjected to iterative training for selection of the optimal feature space among extended 210-sized set, embodying interactive second-order effects between 20 independent features. The optimization process minimizes at equal weight the false positives in SVB-class and false negatives in VB-class. The training with European ST-T, AHA, MIT-BIH Supraventricular Arrhythmia databases found the best performance settings of all classification models: Cluster (30 features), Fuzzy (72 features), LDA (142 coefficients), CT (221 decision nodes) with top-3 best scored features: normalized current RR-interval, higher/lower frequency content ratio, beat-to-template correlation. Unbiased test-validation with MIT-BIH Arrhythmia database rates the classifiers in descending order of their specificity for SVB-class: CT (99.9%), LDA (99.6%), Cluster (99.5%), Fuzzy (99.4%); sensitivity for ventricular ectopic beats as part from VB-class (commonly reported in published beat-classification studies): CT (96.7%), Fuzzy (94.4%), LDA (94.2%), Cluster (92.4%); positive predictivity: CT (99.2%), Cluster (93.6%), LDA (93.0%), Fuzzy (92.4%). CT has superior accuracy by 0.3–6.8% points, with the advantage for easy model complexity configuration by pruning the tree consisted of easy interpretable ‘if-then’ rules.

Highlights

Detection of cardiac arrhythmias is potentially life-saving as any disturbance in the rate, regularity, site of origin or conduction of electrical impulses through the myocardium might refer to a structural or functional heart disease with the risk of developing heart failure [1]
The objective of this work is to develop and compare the best performance of four independent realizations of linear programming beat classifiers based on cluster analysis, linear discriminant analysis (LDA), fuzzy analysis and classification tree (CT), all subjected to iterative training for selection of the optimal feature space
The heartbeat classification performance of Stage 1 and Stage 2 are estimated by three statistical indices that are adopted in the research community to provide comprehensive assessment of imbalanced learning problems [48]: sensitivity (Se), specificity (Sp), positive predictive value doi:10.1371/journal.pone.0140123.g002 (PPV)

Summary

Introduction

Detection of cardiac arrhythmias is potentially life-saving as any disturbance in the rate, regularity, site of origin or conduction of electrical impulses through the myocardium might refer to a structural or functional heart disease with the risk of developing heart failure [1]. Considering the inter-patient and intra-patient variation of the ECG waveform, some beat classification systems aim to improve their performance by taking the advantage from a local expert assistance for initial annotation of a group of typical or pathological beats in one ECG recording rather than only relying on a global learning strategy [2,3,4,5,6,7]. The analysis of the P-QRS-T waveform complexity and regularity of the cardiac cycle duration (RR-interval) is used for extraction of a diverse set of features which are subjected to optimization in different decision support systems aiming at the most reliable classification of normal or abnormal beats with supraventricular or ventricular origin. The dynamic ECG features estimated as a variation of the neighboring RR-intervals are considered in almost all published works

Objectives

Methods

Results

Discussion

Conclusion