Abstract: In order to avoid interference of language meaning on English oral translation results and improve the accuracy of English oral translation, an accurate English oral translation system based on attribute features is proposed. The system extracts oral semantic feature parameters by analyzing input data variables. Semantic feature parameters can capture differences between vocabulary and expression methods, improving translation accuracy. Variational autoencoders is used to capture effective information of features and obtain English spoken semantic matching results. Oral semantic matching is encoded and decoded, and translation rules are set based on parameters to identify interpreting ambiguous word parameters, and CBOW(Continuous Bag-of-Words) model is used to identify and evaluate parameters. Translation rules for complex sentence structures are established and they are connected through semantic translation based on these rules, forming accurate English oral translation results. The experimental results show that the designed English oral translation system has high translation accuracy and can achieve accurate English oral translation. Therefore, it indicates that the studied translation system can meet the requirements of high-precision English oral translation.

Key words: attribute characteristics, english oral translation, continuous bag-of-words ( CBOW) model, semantic features, variational autoencoder, group translation rules