Skip to main content

Table 4 Time to train and apply HMM models on the i2b2/VA and JNLPBA corpora 1

From: Detecting concept mentions in biomedical text using hidden Markov model: multiple concept types at once or one at a time?

 

A set of types considered

Training (sec)

Application (sec)

i2b2

Problem, Test, Treatment

619

42

Problem, Treatment

763

41

Problem, Test

879

42

Problem

1,117

43

JNLPBA

Protein, DNA, Cell Type, Cell line, RNA

3,010

88

Protein, DNA, Cell Type, Cell line

3,812

92

Protein, DNA, Cell Type

4,292

98

Protein, RNA

4,694

100

Protein

4,763

98

  1. 1The experiments were conducted on a server with six-core AMD Opteron 2.8 GHz processors running CentOS 2.6. The reported times are the average of ten runs in ten-fold cross-validation.