Improve metric for candidate selection #40
Labels
enhancement
New feature or request
help wanted
Extra attention is needed
wontfix
This will not be worked on
At present we are using edit distance (Levenshtein distance) which is provided by editdistance package. For many languages (Japanese, Chinese and others) it is not a good similarity score. It was suggested to use strokes score does it have available material to read and implement or any open source package?
The text was updated successfully, but these errors were encountered: