Design of Internet Platform for English Teaching in Colleges and Universities under the Background of Big Data

Abstract

Abstract This paper constructs an English teaching platform based on a speech recognition algorithm and Internet technology, adopts a three-tier architecture to design the web application, and combines the server and browser technology to realize the operation of data interaction with the SQL database. A rectangular bandpass filter is selected for narrow-band information extraction by introducing the sinc function to improve the feature extraction efficiency of speech signals at each frequency. CTC is used to obtain input feature-to-state mapping to solve the sequence problem in speech recognition and achieve end-to-end speech recognition. To achieve the final audio recognition results, the speech decoding algorithm is employed to select the optimal speech output. The results show that in the listening score, the score of the experimental group is 9, and the score of the control is 6. The completion rate of the topics among students who studied on the English teaching platform was higher than that among students who were taught in ordinary English. The completion rate is 0.82 for English questions of grammar type and 0.98 for English questions of listening type, and the average number of passes is the least number of passes, and the number of times is 2.