Classification of Heart conditions by Statistical Characterization of ECG Signal

Abstract

Electrocardiogram (ECG) signal exhibits
important distinctive feature for different cardiac issues.
Automatic classification of electrocardiogram (ECG) signal can
be used for primary detection of various heart conditions.
Information about heart and ischemic changes of heart may be
obtained from cleaned ECG signals. ECG signal has an
important role in monitoring and diacritic of the heart patients.
An accurate ECG classification is challenging problem. The
accuracy often depends on proper selection of observing
parameters as well as detection algorithms. Heart disorder means
abnormal rhythm or abnormalities present in the heart. In this
research work, we have developed a decision tree based
algorithm to classify heart problems by utilizing the statistical
signal characteristic (SSC) of an ECG signal. The proposed
model has been tested with real ECG signal to successfully (60-
98%) detect normal, apnea and ventricular tachyarrhythmia
condition.