文章基本信息

标题：ANALYSIS WITH VARYING QUESTION SIZE IN QUESTION PAPER TRANSLATION
本地全文：下载
作者：SHWETA VIKRAM ; SANJAY K. DWIVEDI
期刊名称：Journal of Theoretical and Applied Information Technology
印刷版ISSN：1992-8645
电子版ISSN：1817-3195
出版年度：2019
卷号：97
期号：15
页码：4011-4020
出版社：Journal of Theoretical and Applied
摘要：Word Sense Disambiguation in question paper translation is a challenging task. Some words in the question sentence can make the entire sentence ambiguous. WSD is a process to remove the ambiguity in a natural sentence to provide the correct sense of the word according to the sentence/context. Works have been done in question answering system to deal with ambiguity, however there has not been much work in resolving ambiguity related issues specially when it comes to translate questions rather than simple text. This paper specially highlights issues in the translation of Wh-questions from English to Hindi. We used five translators to show the impact of translation of Wh-questions using these translations. The experimental analysis of some English questions classified in three categories based on the number of words in each question. After analyzing these translations through MT tools for the three categories of questions, we found that the performance of translations of small questions is much better than that of other category questions having size medium to large. Further the average BLEU score (for all categories) has been found 0.483 for Babelfish which is best whereas Babylon performed poorly with 0.429.
关键词：Machine Translation; Word Sense Disambiguation; Questions; English and Hindi