Closed koheiw closed 2 years ago
Merging #45 (745af45) into master (6a89b78) will decrease coverage by
0.02%
. The diff coverage is100.00%
.:exclamation: Current head 745af45 differs from pull request most recent head 8dcdc37. Consider uploading reports for the commit 8dcdc37 to get more accurate results
@@ Coverage Diff @@
## master #45 +/- ##
==========================================
- Coverage 82.22% 82.20% -0.03%
==========================================
Files 16 16
Lines 1176 1152 -24
==========================================
- Hits 967 947 -20
+ Misses 209 205 -4
Impacted Files | Coverage Δ | |
---|---|---|
R/textstat_simil.R | 97.80% <100.00%> (+1.68%) |
:arrow_up: |
Continue to review full report at Codecov.
Legend - Click here to learn more
Δ = absolute <relative> (impact)
,ø = not affected
,? = missing data
Powered by Codecov. Last update 6a89b78...8dcdc37. Read the comment docs.
Use the
use_nan
argument that was added to proxyC recently. It's functions return the NaN for the empty or uniform vectors in the same way asproxy::simil()
orstats::dist()
, but it is only when measures are correlation or cosine.proxyC returns zero when measures are based on set-theory or vector space model, because common elements of empty sets are zero and distance between origins are zero. Cosine is also a spatial measure but
proxyC::simil()
still returns NaN for empty vectors because its is about angles, not coordinates.