A Multiple-Classifier Framework for Parkinson’s Disease Detection Based on Various Vocal Tests

Table 4

Prediction ability of each vocal test, based on their obtained classification accuracy.

Vocal test ID	Classification accuracy (%)
	-NN; ()	-NN; ()	-NN; ()	-NN; ()	SVM; (linear kernel)	SVM; (RBF kernel)	Naïve Bayes	DA	Mean accuracy ± standard deviation
1	42.5	35	27.5	27.5	47.5	25	52.5	37.5	36.9 ± 10
2	57.5	67.5	70	70	62.5	62.5	60	70	65 ± 5
3	40	42.5	60	55	27.5	50	50	47.5	46.56 ± 10
4	67.5	70	75	67.5	60	65	65	62.5	66.6 ± 4.6
5	67.5	57.5	60	65	67.5	67.5	67.5	65	64.7 ± 3.9
6	62.5	67.5	67.5	72.5	62.5	72.5	72.5	65	67.8 ± 4.3
7	52.5	60	57.5	55	67.5	67.5	70	50	60 ± 7.6
8	57.5	62.5	62.5	70	65	67.5	65	62.5	64 ± 3.8
9	47.5	62.5	65	50	60	50	55	57.5	55.9 ± 6.4
10	62.5	62.5	65	55	55	55	57.5	55	58.4 ± 4.2
11	42.5	60	72.5	72.5	75	72.5	70	72.5	67.2 ± 11
12	50	45	42.5	57.5	65	65	65	65	56.9 ± 9.7
13	40	45	57.5	57.5	52.5	60	60	45	52.2 ± 7.8
14	52.5	60	55	55	50	60	45	55	54.1 ± 5
15	57.5	60	62.5	65	72.5	72.5	65	72.5	65.9 ± 6
16	50	55	62.5	65	72.5	72.5	65	72.5	64.4 ± 8.4
17	60	57.5	57.5	65	72.5	60	72.5	67.5	64.1 ± 6.3
18	45	45	55	65	67.5	67.5	65	67.5	59.7 ± 9.9
19	40	37.5	45	35	40	40	22.5	42.5	37.8 ± 6.9
20	52.5	55	55	47.5	60	62.5	65	62.5	57.5 ± 6
21	47.5	40	40	27.5	45	55	52.5	52.5	45 ± 9
22	42.5	35	52.5	52.5	67.5	55	55	65	53.1 ± 10.7
23	47.5	55	55	47.5	62.5	60	60	50	54.7 ± 5.9
24	35	35	22.5	35	42.5	42.5	37.5	50	37.5 ± 8
25	62.5	62.5	67.5	67.5	62.5	67.5	65	62.5	64.7 ± 2.5
26	55	62.5	57.5	62.5	57.5	57.5	55	55	57.8 ± 3.1