Skip to main content
. 2010 Oct 11;1:24. doi: 10.4103/2153-3539.71065

Table 7.

Performance with improved lexicon. Posttest performance for each frame building method with lexicon improvement

Dynamic-window
ConText
Original test With improved lexicon Original test With improved lexicon
Accuracy 0.962 0.972 0.962 0.980
Precision 0.893 0.898 0.877 0.919
Recall 0.944 0.982 0.961 1.000
F-measure 0.918 0.938 0.916 0.957
Avg kappa (ref std) 0.928