关键词: 文本数据集, 计算人文, 高质量数据集, AI就绪数据集, AI4S

Abstract: [Purpose/Significance] Addressing the problems of diverse forms but unclear quality standards and limited application adaptability of text dataset in computational humanities research, this study focuses the construction of high-quality text datasets and analyzes their connotative characteristics, implementation methods, and application orientations, aiming to identify the development pathways of high-quality text datasets and promote their evolution towards an AI-ready state. [Design/Methodology] This study systematically examines the intrinsic characteristics and construction pathways of high-quality text datasets in computational humanities from the perspectives of concepts, methods, and applications. Furthermore, it evaluates the application advantages and development directions based on the computational capabilities of different forms of text datasets. [Findings/Conclusion] Different types of text datasets exhibit distinct development priorities and application orientations. The "high quality" of these Datasets is not solely determined by the data form or scale,but rather is a comprehensive reflection of both data quality and application adaptability. [Originality/Value] Exploring the coordinated evolution process in the concepts, methods and applications of building high-quality text datasets in computational humanities helps to promote new paradigms in humanities research and provides a systematic data and knowledge foundation for cultural representation and value construction in intelligent models.

Keywords: Text dataset, Computational humanities, High-quality dataset, AI-ready dataset, AI4S