Construction and Optimization of Bilingual Translation Corpus Based on GLR Algorithm

When computers and other machines are used to translate English content, it often happens that a word is translated incorrectly or the context is not smooth due to the improper recognition of the word content and context, which is one of the shortcomings of traditional machine English translation. In order to ensure the accuracy of machine translation (MT), this paper puts forward a multi-outlet analysis table structure when applying GLR algorithm to shallow syntactic analysis, which reduces the complexity of the algorithm, collects and processes language signals, and extracts feature parameters to realize MT intelligent proofreading. The designed MT intelligent proofreading model is analyzed experimentally and the experimental data is recorded. The simulation results show that the proposed GLR algorithm has higher prediction accuracy and recall than the traditional translation proofreading model, which are 19.65% and 18.77% higher than the contrast BPNN. Therefore, it is feasible to apply this model to the construction and optimization of bilingual translation corpus. This algorithm can play a stable role in the construction and optimization of bilingual translation corpus.