Skip to main content

Table 2 Sensitivity and specificity results of various classifiers trained on the hepatobiliary and the acute renal failure data sets, with and without set expansion

From: Improving sensitivity of machine learning methods for automated case identification from free-text electronic medical records

 

Set

Imbalance

SVM

C4.5

MyC

RIPPER

Data set

expansion

ratio

Sens

Spec

Sens

Spec

Sens

Spec

Sens

Spec

Hepatobiliary

No

0.5

0.99

0.03

0.99

0.03

0.99

0.07

0.99

0.04

 

Yes

42

0.89

0.77

0.90

0.79

0.92

0.69

0.91

0.71

Acute renal failure

No

16

0.62

0.92

0.69

0.88

0.69

0.90

0.71

0.89

 

Yes

1363

0.39

0.98

-

-

0.45

0.99

0.41

0.98