Named entity translation with web mining and transliteration

Abstract

This paper presents a novel approach to improve the named entity translation by combining a transliteration approach with web mining, using web information as a source to complement transliteration, and using transliteration information to guide and enhance web mining. A Maximum Entropy model is employed to rank translation candidates by combining pronunciation similarity and bilingual contextual co-occurrence. Experimental results show that our approach effectively improves the precision and recall of the named entity translation by a large margin. 1