I am looking for a way to find out how complicated a language is in terms of having more graphemic similarity between its lexicons.
You can use Levenshtein Distance, Jaro-Winkler Distance or Jaccard Similarity for that. You also find these measures ready implemented for most programing languages (e.g. Jellyfish for python or DKPro Similarity for Java)
Thank you for the answer but these distance metrics won't diffrenciate these two examples:
EcoleÉcole vs. helpheld
Both cases have the same distance with these simple distance metrics. I was looking for a method to have less distance to the first one rather than the second.
if you are using python, it seems a bit easy. you can convert accented alphabets into ascii character by using the module "unicodedata", e.g.,
import unicodedata
s=u'École'
s1=unicodedata.normalize('NFD', s).encode('ascii', 'ignore')
s1 will be "Ecole"
If you want keep the difference, you can encode back with unicode and check which character has been changed.
- Badges
- Science topic
- Similar topics
- Phytochemistry
- Extracts