Date of Publication :30th November 2023

Abstract: In this paper, machine translation of Chinese sentences to Hindi has been implemented using Moses which is a statistical machine translation system. Moses has been trained by using a small self-generated corpora of 500 sentence pair of Chinese-Hindi. This model has been tuned using 100 different sentences of Chinese language. All 500 sentences of both Chinese and Hindi language have been pre-processed before they are used for training of the model. These steps include tokenization, truecasing and cleaning of the corpus. The model has been tested for 35 sentences of Chinese which are translated to Hindi. Further, the same sentences of Chinese are also tested using online Google translator and Microsoft Bing translator. Finally, BLEU score has been calculated for each of the translators and their results are compared. The results of Chinese-Hindi sentence pairs show an accuracy of 67%. Also it gives a BLEU score of 0.5. The accuracy obtained by Google translator and Microsoft Bing are 62% and 36%, respectively.

Reference :