Percentage of part of speech differences between source-side and reference-side spans. ¬EXT is the percentage of span pairs whose POS tags did not exactly match. Since an exact match would be precluded in the case of differing span lengths we also include Jaccard distance11 (JCD) and weighted Jaccard distance12 (WJCD), the latter of which is sensitive to tag frequency. In row one the reference-side spans are produced by the alignment model whereas in row two the analysis uses gold manually annotated span labels.
Comparison . | ¬EXT . | JCD . | WJCD . |
---|---|---|---|
Source vs. Model | 31.08 | 28.94 | 29.31 |
Source vs. Gold | 34.54 | 30.60 | 31.05 |
Comparison . | ¬EXT . | JCD . | WJCD . |
---|---|---|---|
Source vs. Model | 31.08 | 28.94 | 29.31 |
Source vs. Gold | 34.54 | 30.60 | 31.05 |