Construction of Mizo – English Parallel Corpus for Machine Translation

Parallel corpus is a key component of statistical and Neural Machine Translation (NMT). While most research focuses on machine translation, creation studies are limited for many languages no paper Mizo–English exists yet. A high-quality parallel required Natural Language Processing (NLP) activities including Chatbots, Transliteration, Cross-Language Information Retrieval. This work aims to investigate techniques apply them the Mizo-English language pair. Another goal test translation newly constructed corpus. We contributed LF Aligner tool support Mizo sentence alignment in development. Our effort created first large-scale with over 529K sentences. The pre-processed was used Mizo-to-English NMT. It evaluated using BLEU, ChrF, TER scores. system achieved BLEU 45.08, ChrF 65.36, 41.16, setting new benchmark translation.