Standardized Evaluation Method of Pronunciation Teaching Based on Deep Learning

With the advancement of globalization, an increasing number people are learning and using a common language as tool for international communication. However, there clear distinctions between native target language, especially in pronunciation, domestic environment is far from ideal, with few competent teachers. In addition, such cannot achieve computer-assisted (CALL) technology. The efficient combination computer technology teaching methods provides new solution to this problem. core CALL speech recognition (SR) evaluation development deep (DL) has greatly promoted recognition. pronunciation resource collected Chinese college students, whose majors education or who planning obtain better shall be research object paper. study applies standard but builds model based on belief network (DBN). On basis, work improves traditional quality method, comprehensively considers intonation, speaking speed, rhythm, other multi-parameter indicators their weights, establishes reasonable model. systematic results show that article theoretical practical value field phonetics education.