Computer-Assisted English Presentation Learning System Based on Word Stress in Pronunciation

Non-English speaking students are often required to give presentations in English. Using CALL (Computer-Assisted Language Learning), students can practice English pronunciation as many times as they want. Some current CALLs make use of machine learning and these help students to learn English objectively and efficiently. In this paper, we propose and evaluate an English pronunciation training system. The proposed system focuses on improving sentence intonation by paying attention to word stress during pronunciation and aims to make the student's pronunciation of each word similar to that of the model that the student likes. For this purpose, we have developed a system that evaluates the word stress of the student's pronunciation and gives advice on how to make the student's pronunciation close to that of the model.