Romagi of Shan Language to English Using Phrase-Based Statistical Machine Translation Model

Machine translation methods now offer top results for certain foreign languages. Machine Translation is also called the automatic translation process which uses a computerized system to convert one natural language into another. There are numerous studies using machine translation software for both foreign languages and ethnic languages of Myanmar (lower source languages). This article offers a paradigm for translating Shan (Romagi) phrases into English along with phrase-based analysis. The Moses toolkit is suggested for use with the phrase-based statistical machine translation (PBSMT) system for language pairs of Shan (Romagi) and English. GIZA++ was used to align the word-segmented source language and word-segmented destination language. For training, the proposed system uses 2-gram, 3-gram, and 5-gram language models to translate Romagi to English, this system uses KenLM and SRILM. The Moses decoder was used to decode the data and MERT applied the decoder's parameters for the fine-tuning process. Finally, BLEU scores are used to measure and compare the experimental outcomes of the system.