TCM-SD
收藏资源简介:
TCM-SD数据集是首个公开的大规模证型辨识基准,由北京理工大学计算机科学与技术学院创建。该数据集包含54,152条真实世界的临床记录,涵盖148种证型,旨在通过自然语言处理技术推动中医理论的实证研究。数据集的创建过程涉及从网站爬取大量未标记文本,构建中医领域特定的预训练语言模型ZY-BERT。TCM-SD数据集的应用领域主要集中在中医诊断与治疗系统中,特别是证型辨识任务,以解决中医领域数据驱动AI技术的进一步发展需求。
The TCM-SD dataset is the first publicly accessible large-scale benchmark for traditional Chinese medicine (TCM) syndrome differentiation, developed by the School of Computer Science and Technology, Beijing Institute of Technology. It contains 54,152 real-world clinical records spanning 148 syndrome types, with the goal of advancing empirical research on TCM theories using natural language processing (NLP) technologies. The development process of the dataset included crawling a large volume of unlabeled texts from web sources and constructing the TCM domain-specific pre-trained language model ZY-BERT. Primarily, the TCM-SD dataset is utilized in TCM diagnosis and treatment systems, particularly for syndrome differentiation tasks, to address the demand for further advancement of data-driven AI technologies within the TCM domain.

- 1TCM-SD: A Benchmark for Probing Syndrome Differentiation via Natural Language Processing北京理工大学计算机科学与技术学院 · 2022年



