Construction of modern Chinese standard for computer language disfluency detection
Jiantao Li, Yunqiu Zhang, Jianshe Zhou, Jie Liu · 7th International Symposium on Advances in Electrical, Electronics, and Computer Engineering · 2022
Spoken text disfluency detection is an important component of speech computer recognition systems, and its goal is to effectively identify and remove spoken phenomena such as repetition, pauses, corrections and redundancy contained in AI speech recognition text data, thereby making spoken text data more concise and increasing the readability of its text data, and the technology helps to improve the correctness of computer language information processing tasks. However, in computer science or in linguistics, research on computer language disfluency detection techniques for modern Chinese word classes is almost in an academic gap. Based on a qualitative and quantitative analysis of 50,000 items of Chinese conference corpus, this study defines what is non-sentence elements in modern Chinesen and using the spoken language corpus, established a reference model for computer processing of disfluency detection components.