基于对话信息的口语语料库应用研究 ——属性信息与多模态信息的活用方法与参考价值

文章以收录了多模态信息的口语语料库作为研究对象，对成熟的日语口语语料库进行了详细的分析。通过对对话中“笑声”的出现方式及特征进行实例分析，结合属性信息、对话特征信息、语调变化、沉默等数据，对口语语料库中多种信息的相互关联及活用方法进行了探讨。在此过程中，对口语语料库中多模态信息的科研价值进行了验证。 

The Application of a Spoken Language Corpus Based on Information Embodied in Conversations: Practical Use and Evaluation of Attributes and Multimodal Information

This study employs a Japanese spoken language corpus containing multimodal information to organize and summarize relevant data.An empirical study using attributes such as “laugh,” “rising intonation,” “silence,” and “backchannel” is conducted.A comprehensive analysis of the interaction and practical use of information in multiple forms is implemented.Throughout the process,the value of multimodal information embodied in spoken language corpus is confirmed.