Skip to main content

Table 4 Classification results for suicidal ideation based on acoustic and linguistic features

From: Acoustic and language analysis of speech for suicidal ideation among US veterans

Feature set Model Specificity
(std)
Sensitivity (std) PPV
(std)
NPV
(std)
Accuracy
(std)
AUC
(std)
F1
(std)
Acoustic KNN 0.52 (0.23) 0.78 (0.13) 0.21 (0.03) 0.94 (0.06) 0.55 (0.19) 0.69 (0.11) 0.33 (0.07)
LR 0.64 (0.15) 0.78 (0.11) 0.26 (0.02) 0.95 (0.05) 0.66 (0.13) 0.78 (0.12) 0.38 (0.08)
RF 0.64 (0.12) 0.76 (0.17) 0.25 (0.03) 0.94 (0.05) 0.65 (0.09) 0.76 (0.06) 0.36 (0.04)
SVM 0.54 (0.30) 0.74 (0.18) 0.21 (0.04) 0.93 (0.07) 0.56 (0.25) 0.63 (0.25) 0.33 (0.10)
XGB 0.74 (0.21) 0.67 (0.18) 0.29 (0.04) 0.93 (0.06) 0.73 (0.17) 0.77 (0.08) 0.43 (0.12)
DNN 0.70 (0.06) 0.66 (0.20) 0.26 (0.03) 0.93 (0.03) 0.70 (0.05) 0.75 (0.06) 0.35 (0.08)
Linguistic KNN 0.69 (0.15) 0.38 (0.14) 0.16 (0.03) 0.87 (0.05) 0.65 (0.11) 0.52 (0.07) 0.22 (0.03)
LR 0.57 (0.24) 0.66 (0.20) 0.20 (0.05) 0.91 (0.07) 0.58 (0.18) 0.62 (0.06) 0.29 (0.04)
RF 0.64 (0.17) 0.76 (0.09) 0.26 (0.09) 0.94 (0.08) 0.65 (0.14) 0.74 (0.07) 0.38 (0.09)
SVM 0.63 (0.33) 0.48 (0.40) 0.13 (0.09) 0.88 (0.08) 0.61 (0.24) 0.49 (0.15) 0.19 (0.14)
XGB 0.70 (0.16) 0.67 (0.17) 0.27 (0.08) 0.93 (0.05) 0.69 (0.13) 0.72 (0.04) 0.37 (0.06)
DNN 0.50 (0.15) 0.66 (0.25) 0.16 (0.04) 0.90 (0.06) 0.52 (0.11) 0.63 (0.14) 0.25 (0.07)
Acoustic and Linguistic KNN 0.61 (0.22) 0.78 (0.22) 0.26 (0.09) 0.94 (0.06) 0.63 (0.19) 0.69 (0.15) 0.37 (0.11)
LR 0.74 (0.14) 0.78 (0.19) 0.35 (0.14) 0.95 (0.05) 0.75 (0.11) 0.77 (0.12) 0.46 (0.13)
RF 0.70 (0.16) 0.84 (0.09) 0.32 (0.08) 0.96 (0.08) 0.72 (0.13) 0.80 (0.06) 0.45 (0.09)
SVM 0.86 (0.13) 0.52 (0.32) 0.32 (0.21) 0.92 (0.05) 0.82 (0.09) 0.64 (0.27) 0.38 (0.24)
XGB 0.79 (0.11) 0.74 (0.18) 0.37 (0.11) 0.95 (0.05) 0.78 (0.08) 0.77 (0.05) 0.47 (0.07)
DNN 0.68 (0.06) 0.70 (0.15) 0.24 (0.04) 0.93 (0.03) 0.68 (0.04) 0.77 (0.08) 0.36 (0.05)
\