Skip to main content

Table 4 Regression results with feature selection using the clustering approach with 2, 9, and 15 clusters. Turkish database is used. Statistically significant (p<0.05) improvement is shown in underlined bold

From: Depression-level assessment from multi-lingual conversational speech data using acoustic and text features

Dim

Baseline

2-Cluster

9-Cluster

15-Cluster

5

12.56

11.35

13.14

11.99

10

12.45

10.95

13.42

12.25

15

12.08

11.13

13.07

11.75

20

12.87

11.74

13.23

12.95

40

13.28

12.33

13.73

12.06

80

11.58

12.72

13.33

10.83

100

11.75

13.22

13.09

10.97

200

11.32

11.72

12.66

11.50

400

11.42

11.83

12.00

11.40

800

11.31

11.62

11.70

11.64