Translation of English-Chinese person name based on dictionary, bilingual corpus and web mining

Abstract

English person names are translated into Chinese person names using the combined method-dictionary, entropy alignment model and web mining. Entropy alignment model makes use of the dictionary of person names and surnames, bidirectional conditional probability and transliteration similarity. Web mining makes use of rules, clue words, transliteration similarity and conditional probability. The experimental results show word name dictionary combined with entropy alignment model can achieve high precision and recall rate for large scale of parallel corpus. Web mining helps to improve the precision of name translation especially for those wrong alignments and nonaligned names of entropy alignment model.