Corpus-based Research on English-Chinese Translation Teaching Combining Vocabulary Learning and Practice
Nan Zhang
DOI: https://doi.org/10.52783/jes.3187
2024-04-04
Abstract:As the global economy is developing, real time translation among major influential languages such as English, Chinese has become a necessity. Translation from English to Chinese is difficult since Chinese has a different grammar and ambiguous word boundaries. Furthermore, there are issues with the current Chinese-English machine translation, including difficult-to-understand extended sentences and inaccurate word translation. To address this drawback, the methodology for more precisely categorizing vocabulary in both English and Chinese is described in this research. The data are collected from UM-Corpus news dataset. Then, using SEGBOT: Neural Text Segmentation method, English words are segmented from the sentences available on the dataset. Afterward, the data are fed to pre-processing. In pre-processing segment; words missing in translation from the news dataset are eliminated and enhances the input data utilizing Iterated Square-Root Cubature Kalman Filter. The outcome from the pre-processing data is transferred to the Pyramidal Convolution Shuffle Attention Neural Network (PCSANN). The word order, grammatical structure, consistent style and smooth flow for English to Chinese translation are successfully classified by using PCSANN. The Waterwheel Plant Algorithm (WWPA) is used to optimize the weight parameter of PCSANN. The proposed PCSANN-WWPA is applied in python working platform. Performance metrics, like accuracy, precision, F1-score, and recall are examined to compute proposed method. The gained results of the proposed PCSANN-WWPA method attains higher accuracy of 16.65%, 18.85%, and 17.89%, higher sensitivity of 16.34%, 12.23%, and 18.54% and higher precision of 14.89%, 16.89%, and 18.23%. The proposed ECTT-PCSANN-WWPA method is compared with the existing methods such as ECTT-RNN, ECTT-DNN, and ECTT-MLPNN models respectively.