1Ph. D. Research, Scholar, IKG Punjab Technical University, Kapurthala, Punjab, India
2Department of Computer Science, Punjabi University, Patiala, Punjab, India
3Department of Applied Sciences, MIMIT, Malout, Punjab, India
Online published on 4 May, 2019.
Parallel corpus is the key resource for English Punjabi machine translation. At wide level there is no availability of English-Punjabi Corpora. There is a primary requirement of parallel corpus for the training of statistical machine translation.
In this paper, our work focuses on building English-Punjabi corpus at large scale. It posed difficulties and the intensive labor to develop the corpus. We are intricate on the collection as well as the flow of work for the construction of parallel corpus. Now after getting the raw text, we need to refine the corpus in such a way that every source language sentence should have corresponding target language sentence.
The paper attempts to explore existing tools as well as building new tools. One of the goals is alignment of bilingual corpus. The alignment algorithms are used to tune the sentences. The accuracy depends on the type of corpus.
A cautious endeavor has been made to capture different types of texts.
Bilingual corpora, Machine-translation, English, Punjabi, NLP
