Skip to main content

Table 2 Improvement of minimizing the negative effects caused by the differences

From: Investigating heterogeneous protein annotations toward cross-corpora utilization

Difference Strategy A A+GENIA G G+GENIA
- - 77.68 66.16 69.65 63.62
boundary of annotated entities to loosen matching criterion 85.20 80.23 84.11 77.33
annotated entities of interest to find compatible annotations   85.21   82.17
sentence selection to select compatible sentences only   85.96   83.75
  1. The scores shown in this table are the F-scores of training with the following data: A(IMed), A(IMed)+GENIA, G(ENETAG) and G(ENETAG)+GENIA.