Key result
An automatic classification method combining heart-rate, Mahalanobis distance, and Legendre polynomial features achieved an average sensitivity of 98.1% and specificity of 85.2% for detecting true ischaemic episodes.
Population
Ambulatory ECG records from the Long-Term ST Database containing ischaemic and non-ischaemic heart-rate…
Authors
Loading...
May aid automated distinction of ischaemic ST changes; leaves open prospective validation before clinical adoption.
An automated decision tree-based classification method using specific ECG morphological features can accurately distinguish between ischaemic and non-ischaemic heart-rate related ST segment deviations.
Faganeli et al. (2010) studied Transient ischaemic and non-ischaemic heart-rate related ST segment deviation episodes. Automatic classification using decision trees with heart-rate, Mahalanobis distance, and Legendre polynomial features was evaluated on Classification performance (sensitivity and specificity). An automatic classification method combining heart-rate, Mahalanobis distance, and Legendre polynomial features achieved an average sensitivity of 98.1% and specificity of 85.2% for detecting true ischaemic episodes.
Synapse has enriched one closely related paper. Consider it for comparative context: