A Spoken English Teaching System Based on Speech Recognition and Machine Learning

The learning model and environment are two major constraints on spoken English by Chinese learners. maturity of computer-aided language brings a new opportunity to Based speech recognition machine learning, this paper designs teaching system, determines the overall architecture functional modules system according system’s demand. Specifically, MATLAB was adopted realize recognition, generate module. Combined with algorithm, deep belief network (DBN)-support vector (SVM) proposed classify detect errors in pronunciation; module also scores quality corrects pronunciation. This extended evaluation created. Next, several experiments were carried out test multiple attributes including accuracy pronunciation classification error detection, rates different environments vocabularies, real-timeliness recognition. results show that our achieved good performance, realized preset design goals, satisfied user research provides an important theoretical practical reference transforming method, improving