A Hyper-Solution Framework for SVM Classification: Application for Predicting Destabilizations in Chronic Heart Failure Patients

Table 1: Performance Indexes for the Three Framework Implementations

	Balanced Classification Accuracy^*	Sensitivity	False Positive Rate	Time
TS-based framework	85.56%	81.82%	10.69%	2:51:44
GA-based SVM framework (with nominal genes)	87.35%	90.91%	16.21%	1:42:40
GA-based SVM framework (with mixed genes)	86.49%	90.91%	17.93%	3:02:39

^* Performance measure for optimization (fitness and benefit of solutions, for GA e TS respectively).