Skip to main content

Table 4 Performance of the SVM-based system for different feature combinations in 10-fold cross-validation

From: Recognition of medication information from discharge summaries using ensembles of classifiers

Words

History

Morphology

POS

Orthographies

Semantics tag

Pre/Re/F-score

*

     

87.09/77.05/81.76

*

*

    

90.34/78.17/83.81

*

 

*

   

91.81/80.74/85.91

*

  

*

  

89.92/78.54/83.84

*

   

*

 

91.62/81.32/86.15

*

    

*

92.38/86.73/89.47

*

*

*

   

91.72/81.08/86.06

*

*

*

*

  

91.81/81.06/86.10

*

*

*

*

*

 

91.78/81.29/86.22

*

*

*

*

*

*

93.75/87.55/90.54

  1. * indicates the corresponding feature in the first row is used. Rows indicate the combination of features and their micro-averaged Precision(Pre)/Recall(Re)/F-score. The best F-score is underlined. F-scores that are significantly higher (p-values < 0.05) than all F-scores from rows above are shown in bold.
Back to article page

BMC Medical Informatics and Decision Making

ISSN: 1472-6947

Contact us