Heart Risk Failure Prediction Using a Novel Feature Selection Method for Feature Refinement and Neural Network for Classification

Table 3

Results of different subsets of features for the heart disease dataset.

n						Sen. (%)	Spec. (%)
12	1	2	19	85.99	90.00	85.36	93.87	0.799
12	2	2	64	85.55	90.00	82.92	95.91	0.801
12	4	3	4	85.50	92.22	82.92	100	0.851
12	5	2	2	83.09	91.11	87.80	93.87	0.820
12	6	2	35	84.05	91.11	85.36	95.91	0.822
12	7	6	8	85.99	90.00	90.24	89.79	0.799
12	10	5	66	87.43	91.11	92.68	89.79	0.822
12	11	3	25	88.40	91.11	85.36	95.91	0.822
12	12	2	12	84.54	91.11	90.24	91.83	0.820
11	(5, 6)	50	2	83.57	93.33	85.36	100	0.872
6	(3 to 9)	16	42	93.23	92.22	90.24	93.87	0.843
13	—	2	4	85.02	90.00	87.80	91.83	0.798