|ijפ(Conference Paper)

Correcting Chinese Repetition Repairs in Spontaneous Speech

Yue-Shi Lee and Hsin-Hsi Chen

Department of Computer Science and Information Engineering

National Taiwan University

Taipei, Taiwan, R.O.C.


Disfluencies involving speech repairs pose serious problems for spoken language processing systems. However, which cues in speech signals may facilitate Chinese repair processing is not known. This paper concerns the acoustic and prosodic analysis for correcting Chinese repetition repairs in spontaneous speech. A large spoken corpus is examined in this study. The experimental results show that our method can achieve the precision rate of 93.87% and the recall rate of 90.65%, without using the lexical information.