Prediction of Sudden Cardiac Death Using Ensemble Classifiers

Sudden Cardiac Death (SCD) is a medical problem that responsible for over 300,000 deaths per year in the United States and millions worldwide. SCD defined as death occurring from within one hour of onset acute symptoms, an unwitnessed absence pre-existing progressive circulatory failures or other causes deaths, during attempted resuscitation. due to cardiac reasons leading cause among Congestive Heart Failure (CHF) patients. The use Electronic Medical Records (EMR) systems has made wealth data available research analysis. Supervised machine learning methods have been successfully used diagnosis. Ensemble classifiers are known achieve better prediction accuracy than its constituent base classifiers. In effort understand factors contributing SCD, on 2,521 patients were collected Trial (SCD-HeFT). included 96 features gathered period 5 years. goal this work was develop model could accurately predict based features. Cox proportional hazards score then ExtraTreesClassifier algorithm boosting mechanism create ensemble. We tested system at point 180 days. Our best results 180-days with 0.9624, specificity 0.9915, F1 0.9607.