Changes between Initial Version and Version 1 of Measurements
- Timestamp:
- Mar 17, 2009, 10:41:57 AM (15 years ago)
Legend:
- Unmodified
- Added
- Removed
- Modified
-
Measurements
v1 v1 1 The latest preliminary precision measurements (performed with SET version 0.2) show that the precision of the parser dependency output ranges between 75 and 86 percent with respect to the human-anotated Czech corpus data, depending on the particular testing set.