JPS-DAPRINFO

JPS-DAPRINFO是由大阪大学和先进电信研究所国际合作创建的数据集，专注于日语对话行为分析和人物相关信息检测。该数据集包含20130条标注样本，来源于50个30分钟的日语双人对话访谈。数据集通过参考Stolke等人的方法进行标注，涵盖13种标签，如主观信息、客观信息等。此数据集不仅用于人物相关信息检测，还可用于对话结构分析和文本分类任务，有助于提升人机对话系统的理解和响应能力。

JPS-DAPRINFO is a dataset developed through international collaboration between Osaka University and the Advanced Telecommunications Research Institute International (ATR). It focuses on Japanese conversational behavior analysis and person-related information detection. This dataset contains 20,130 annotated samples sourced from 50 30-minute Japanese two-person conversational interviews. The dataset was annotated following the methodology proposed by Stolke et al., covering 13 categories of labels including subjective information, objective information, and others. In addition to being used for person-related information detection, this dataset can also be applied to conversational structure analysis and text classification tasks, which helps to improve the understanding and response capabilities of human-machine dialogue systems.