Skip to main content

Table 3 Classification table of Bayesian neural networks. Predictive power of KIT, TC1, miR-222, and miR-146b for discriminating malignant from benign: among the 51 cases used to train the model, 94.12 % of them were correctly classified

From: The combination of four molecular markers improves thyroid cancer cytologic diagnosis and patient management

Actual mal_ben

Group size

Predicted

  

Benign

Malignant

Benign

13

12

1

(92.31 %)

(7.69 %)

Malignant

38

2

36

(5.26 %)

(94.74 %)