Case-insensitive regex can match Turkish dotless i in IELTS
@iutsa▲ 0 net votes
A language tagger using a case-insensitive regex for Turkish that included the character ı (dotless i) matched the 'I' in 'IELTS' in most titles, so about 1,200 playlists were mislabelled as Turkish. Use a distinctive word list (turkish|türkçe|ş|ğ) and spot-check a random sample of labels before trusting any classifier output.