An English Pronunciation Evaluation Model Based on Acoustic Phonetics Features

Oral English practice constitutes a crucial component of English language acquisition for non-native speakers; however, traditional classroom teaching methods are increasingly insufficient to address the diverse needs of learners. This paper enhances the evaluation of pronunciation quality within Computer-Assisted Language Learning (CALL) system. To assess the fluency of English learners' pronunciation, this paper introduces the innovative feature known as wordDurationRatio. Additionally, an intonation evaluation method utilizing dynamic time warping (DTW) algorithms is explored. Finally, this study incorporates support vector regression (SVR) algorithms to effectively fuse features, enhancing the assessment result of pronunciation quality.