Risk stratification for Arrhythmic Sudden Cardiac Death in heart failure patients using machine learning techniques

Arrhythmic Sudden Cardiac Death (SCD) is still a major clinical challenge even though much research has been done in the field. Machine learning techniques give a powerful tool for stratifying arrhythmic risk. We analyzed 40 Holter recordings from heart failure patients, 20 of which were characterized as high arrhythmia risk after 16 months follow up. The two groups (high and low risk) were not statistically different in basic clinical characteristics. We performed windowed analysis and computed 25 Heart Rate Variability (HRV) indices. We fed these indices as input to two classifiers: Support Vector Machines (SVM) and Random Forests (RF). The classification results showed that the automatic classification of the two groups of subjects is possible.