CN-EN interactive translation system based on semantic similarity and RNN algorithm

With the deepening of globalisation and the increasingly frequent exchange of information, the need for a Chinese-English interactive translation system is becoming increasingly urgent. The study combines semantic similarity with RNN algorithms to enhance the translation system's capacity to manage semantic complexity. Traditional methods lack accuracy in processing semantic information and struggle with understanding long sentence structures and contextual cues. The innovation combines semantic similarity with the RNN algorithm to create a seamless Chinese-English translation model, enhancing accuracy and fluency. Experimental results demonstrate a reduction in convergence iterations to 8, with a 71.74% decrease in the sum of square errors and a 55.56% increase in relative convergence speed, enabling rapid global optimisation. The algorithm accuracy increased by 31.2% and 6% respectively, and the F value increased by 17.4% and 8.1%. The average translation time per sentence is 5.8 milliseconds, meeting real-time performance requirements.