Title: Application of machine translation in bilingual knowledge management

Authors: Mahmudul Sheikh; Sumali Conlon

Addresses: Rust College, Holly Spring, MS 38635, USA ' School of Business Administration, University of Mississippi, University, MS 38677, USA

Abstract: International usage of information systems demands multilingual support to the creation and management of user oriented knowledge. Human translation for every pair of natural language is not a viable option. This study investigates the applicability of a machine translator to provide cross-lingual support to the questions and answers (QnA) of an online system. As a test case, we developed a domain specific English to Bengali translator named DENGALI to examine the applicability of a domain specific machine translator in a QnA setting. DENGALI was developed using a domain specific dictionary and a part-of-speech-based word alignment algorithm. The chosen test domain was the QnA of an online travel guide named 'Fodor'. For this application case, DENGALI was found more effective than Google translator and another well-known English to Bengali translator named ONUBADOK.

Keywords: English to Bengali translation; point-wise mutual information; PMI; statistics; statistical machine translation; SMT; bilingual knowledge management; multilingual support; questions and answers; Q&A; online travel guides; Fodor; domain specific dictionaries; word alignment.

DOI: 10.1504/IJIIM.2013.053794

International Journal of Intercultural Information Management, 2013 Vol.3 No.2, pp.123 - 137

Available online: 08 May 2013 *

Full-text access for editors Access for subscribers Purchase this article Comment on this article