Skip to main content

Table 2 Performances of Classifiers

From: Can natural language processing help differentiate inflammatory intestinal diseases in China? Models applying random forest and convolutional neural network approaches

 

Sensitivity

Specificity

AUC

  

RF

CNN

RF

CNN

RF

CNN

  

UC and CD

0.89

–

0.84

–

0.94

–

  

UC and ITB

0.83

–

0.82

–

0.89

–

  

CD and ITB

0.72

0.90

0.77

0.77

0.82

0.91

  
 

Precision

Recall

F1 score

Accuracy

RF

CNN

RF

CNN

RF

CNN

RF

CNN

UC

0.97

0.99

0.97

0.97

0.97 ± 0.01

0.98 ± 0.01

  

CD

0.65

0.87

0.53

0.83

0.58 ± 0.02

0.85 ± 0.01

  

ITB

0.68

0.52

0.76

0.81

0.72 ± 0.02

0.63 ± 0.02

  
       

0.77 ± 0.02

0.88 ± 0.01

  1. AUC Areas under the curve; UC Ulcerative colitis, CD Crohn’s disease, ITB Intestinal tuberculosis, RF Random forest, CNN Convolutional neural network
  2. The confidence interval of accuracy score is calculated by
  3. \( accuracy\pm {z}_{0.975}\sqrt{\frac{accuracy\left(1- accuracy\right)}{n}} \)
  4. The confidence interval of F1 score is estimated by the bootstrap method
Back to article page

BMC Medical Informatics and Decision Making

ISSN: 1472-6947

Contact us