Deep Learning Scoring Model in the Evaluation of Oral English Teaching

Abstract

This study is aimed at improving the accuracy of oral English recognition and proposing evaluation measures with better performance. This work is based on related theories such as deep learning, speech recognition, and oral English practice. As the literature summarized, the recurrent neural network was the calculation standard, and the oral English speech recognition indicators were the main basis on which an English speech recognition model was constructed. Then, 20 English majors and 5 sets of English sentence patterns were randomly selected as the research objects. The correction standards for English oral errors were introduced into the model to achieve further improvement. The research results showed that the average concordance rate of speech recognition reached 91% through the model test. The concordance rates of words, speech, and intonation in recognition were 89%, 91%, and 86%, respectively. The model could be used as an evaluation system for English speech recognition. Therefore, the application of the deep learning scoring model in the evaluation of oral English teaching was researched in this work, which provided an effective basis for the evaluation of intelligent English teaching.