2 ms·
You'd probably want to set your expected % based on word-starting tokens in some large English-language (or Latin-script) corpus, due to the unequal distributio
by nerdponx 4y ago
You'd probably want to set your expected % based on word-starting tokens in some large English-language (or Latin-script) corpus, due to the unequal distribution of letters in English and languages that use Latin script generally.
- bombcar 4y agoAnd then probably balance it somehow by what states have which letters, and the fact that Savings begins with an S. B probably wins because of Bank.