Транслитерация на основе подстрок

We propose an unsupervised approach for substring-based transliteration which in- corporates two new sources of knowledge in the learning process: (i) context by learning substring mappings, as opposed to single …...