Table 3 Performance metrics of classifiers.

From: Taking a look at your speech: identifying diagnostic status and negative symptoms of psychosis using convolutional neural networks

 

AUC

Test accuracy (%)

Precision

Recall

F1 score

Diagnostic classifier

 HC

0.8651

87.83

0.7446

0.8365

0.7879

 SSD

  

0.9366

0.8938

0.9147

Median-split classifier

 Sub-median

0.733

80.46

0.9627

0.8188

0.8849

 Above-median

  

0.2431

0.6473

0.3535

Blunted affect (N1) classifier

 Mildly ill (N1 ≤ 3)

0.7856

87.84

0.9734

0.8936

0.9318

 Severely ill (N1 > 3)

  

0.325

0.6776

0.4393

Back to article page