Ранжирование пар слов

Strictly corpus-based measures of seman- tic distance conflate co-occurrence infor- mation pertaining to the many possible senses of target words. We propose a corpus–thesaurus hybrid method that uses soft constraints …...