What question did this study set out to answer?

April 16, 2026Open Access

Online oral English teaching system based on speech recognition technology and machine learning

Key Points

The aim is to enhance pronunciation error detection and feedback precision in online oral English teaching using advanced technologies.
Designed a system integrating speech recognition and machine learning.
Utilized the MFCC-DBN model with feature fusion for core detection.
Implemented a multi-classifier based on support vector machines (SVM).
Analyzed data from 1,610 expert-annotated phoneme samples.
Achieved high accuracy in detecting both sample-sufficient and small-sample error types.
Outperformed LDA-SVM and Wav2Vec2.0-SVM in accuracy and standard error.
Demonstrated stronger robustness and efficiency with limited data.

Abstract

To enhance the intelligence of pronunciation error detection and feedback precision in online oral English teaching, this study designs a system combining speech recognition and machine learning.Its core detection module uses the MFCC-DBN model with feature fusion, and builds an SVM-based multi-classifier.Experimental data comes from the CSTR VCTK Corpus and the speech accent archive, containing 1,610 expert-annotated phoneme samples.The model yields high accuracy for both sample-sufficient and small-sample error types.Compared with LDA-SVM and Wav2Vec2.0-SVM, it outperforms them in accuracy and standard error.Results prove the fusion model's stronger robustness and efficiency with limited data, offering a practical technical approach to boost learners' cross-cultural communication competence.

Perguntar à IA

Bookmark

View Full Paper