The Performance of Short-Term Heart Rate Variability in the Detection of Congestive Heart Failure

Table 1

Classification results. KNN (-nearest-neighbor) with and without MinMax normalization. KNN/GA (genetic algorithm) optimization with and without MinMax normalization. The results use a 23-fold cross-validation where Ac (accuracy), Se (sensitivity), and Sp (specificity) quantify the performance assessment of the classifier using features.

Algorithm method	With MinMax normalization					Without MinMax normalization
		Ac (%)	Se (%)	Sp (%)	Features	Ac (%)	Se (%)	Sp (%)	Features
KNN	01	92.75	78.26	100.0	16	73.91	78.26	71.73	16
KNN	03	95.65	86.95	100.0	16	82.60	60.87	93.47	16
KNN	05	92.75	78.26	100.0	16	89.85	82.60	93.47	16
KNN	07	92.75	78.26	100.0	16	86.97	73.91	93.47	16
KNN	09	92.75	78.26	100.0	16	91.30	82.60	95.65	16
KNN	11	92.75	78.26	100.0	16	92.75	82.60	97.82	16
KNN	13	92.75	78.26	100.0	16	94.20	86.95	97.82	16
KNN/GA	01	98.55	100.0	97.82	09	92.75	91.30	93.47	03
KNN/GA	03	98.55	95.65	100.0	08	94.20	95.65	93.47	03
KNN/GA	05	100.0	100.0	100.0	05	94.20	91.30	95.65	08
KNN/GA	07	98.55	95.65	100.0	05	95.65	91.30	97.82	07
KNN/GA	09	98.55	95.65	100.0	05	94.20	91.30	95.65	07
KNN/GA	11	98.55	95.65	100.0	05	94.20	86.95	97.82	09
KNN/GA	13	98.55	95.65	100.0	06	94.20	86.95	97.82	07