Abstract

In this study, multidimensional feature extraction is performed on the U-language recordings of the test takers, and these features are evaluated separately, with five categories of features: pronunciation, fluency, vocabulary, grammar, and semantics. A deep neural network model is constructed to model the feature values to obtain the final score. Based on the previous research, this study uses a deep neural network training model instead of linear regression to improve the correlation between model score and expert score. The method of using word frequency for semantic scoring is replaced by the LDA topic model for semantic analysis, which eliminates the need for experts to manually label keywords before scoring and truly automates the critique. Also, this paper introduces text cleaning after speech recognition and deep learning-based speech noise reduction technology in the scoring model, which improves the accuracy of speech recognition and the overall accuracy of the scoring model. Also, innovative applications and improvements are made to key technologies, and the latest technical solutions are integrated and improved. A new open oral grading model is proposed and implemented, and innovations are made in the method of speech feature extraction to improve the dimensionality of open oral grading.

Highlights

  • In recent years, computer-assisted teaching systems have become one of the hot research topics in the fields of computer science and education [1]

  • We study audio processing, speech recognition, automatic essay marking, and deep learning techniques to design and implement a multifeatured intelligent automatic marking model based on the data generated from the language training system of Beijing University of Posts and Telecommunications. e model is designed to solve the problem of automatic marking of open-ended oral English questions and to reduce teachers’ marking pressure

  • Principal Component Analysis (PCA) uses the decomposition of feature bases into orthogonal transformation matrices to convert the original feature vector into a low-dimensional noncorrelated and orthogonal linear feature vector. is new low-dimensional feature vector is determined by the variance of the projection and is ordered from largest to smallest. e first principal component corresponds to the direction with the largest variance and so on, and the last components correspond to the direction with the smallest variance

Read more

Summary

Introduction

Computer-assisted teaching systems have become one of the hot research topics in the fields of computer science and education [1]. With the development of speech recognition technology and the maturity of essay marking systems, it is expected that technically speaking the automatic marking of open-ended speaking questions can be overcome and reaches a practical level [7]. We study audio processing, speech recognition, automatic essay marking, and deep learning techniques to design and implement a multifeatured intelligent automatic marking model based on the data generated from the language training system of Beijing University of Posts and Telecommunications. A malefactor is a malefactor of existing noise reduction techniques, combining traditional noise reduction algorithms with deep learning It applies to speech noise reduction for automatic grading of speaking language. Algorithms that implement these functions are investigated to improve scoring accuracy

Multifeature Fusion Speaking Test Detection Algorithm
Results and Analysis
Conclusion
Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call