Paper Title
Cross-Lingual Name Entity Transliteration System
Abstract
Name Entity Translation has become a major challenge for Machine Translation systems, especially when
languages are from different script. Through this paper, we are proposing a system which aims to convert proper nouns from
one Indian Language into another Indian Language with English as intermediate language. We call the proposed system as
Cross-Lingual Name-Entity transliteration system (CLNET). We have employed the hybrid approach for machine
transliteration of proper nouns from English to Indian Languages and vice-versa. The hybrid approach is a combination of
direct mapping, Gazetteer search, rule based approach. We have implemented approximately 1000+ rules with help of
linguist to improve accuracy of transliteration. Rigorous testing was done on test data covering proper nouns, person names,
location names. The system found to give accuracy of about 80%.
Index Terms— Cross Lingual Information Retrieval, Forward Transliteration, Name Entity Transliteration, Natural
language Processing, Reverse Transliteration.