Phonetic Feature and Pronunciation Improvement for English Learners
Lihui Jiang · 2024
With the development of science and technology, this method has become an efficient method, the combination of sound extraction technology and machine learning methods, through the Gaussian mixture model (GMM), the important features of the speech signal can be captured, the speech features are modeled and the method of improving pronunciation is proposed. To help English learners improve their language characteristics and oral grammar. Through Mel-scale Frequency Cepstral Coefficients (MFCC) combined with GMM, the language characteristics of English learners were extracted and the spoken pronunciation was improved, and the language characteristics and pronunciation of English learners were infinitely close to the standard pronunciation, and the final similarity rate after improvement was as low as 96.3% and as high as 98.1%. MFCC combined with GMM algorithm is of great help for speech feature analysis and pronunciation improvement in English learning.