A multimedia network English listening teaching model based on confidence learning algorithm of speech recognition

Abstract

Language is the most important communication tool of human beings, and listening is one of the basic skills of language expression. Without good listening comprehension ability, it is impossible to use language flexibly to communicate. Due to the influence of traditional education mode, Chinese students' English listening is generally poor. Therefore, a new English listening teaching mode is needed to help students improve their English listening. In this paper, multimedia network technology is used to realize the integrated English listening teaching of listening, speaking and dictation skills, and a multimedia network English listening teaching model based on speech recognition confidence learning algorithm is proposed. In order to improve the effectiveness of the mainstream confidence method based on Lattice posterior probability, this paper proposes an improved confidence algorithm based on Lattice posterior probability, and then converts the obtained confidence score into discriminant confidence score by Support Vector Machine(SVM) to further enhance the discriminant ability of confidence. Aiming at the imbalance of training data, a score correction strategy is proposed. The experiment shows that the English listening teaching model realized by using multimedia network technology can effectively enhance the students interest in learning and improve their listening ability. And the improvement of the mainstream confidence method based on Lattice posteriori probability can effectively improve the recognition ability of the algorithm and further improve the students’ English listening learning effect.