中医药知识表示与推理方法探讨

Research on Knowledge Representation and Reasoning Methods of Chinese Medicine

中医药领域知识有较强的专指性,是一个相对封闭的系统。本文从文献、医案、方药和专病专科等知识库构建方面介绍分析了各领域知识结构特点及其主要应用,分析了中医药领域的知识表示技术及推理方法,认为中医药各资源知识结构特点不一,其模糊性是知识库发展的主要瓶颈;词向量的表示学习方法与基于案例的推理方法在中医药领域有较好的应用前景,从证素单元建模有助于提取深层次特征和模型的泛化。未来应统一部署制定相应的标准规范,以共享共建方式构建领域全局视图,同时注重知识评审和更新机制,提高知识质量。

Knowledge in the field of traditional Chinese medicine has its specificity and is a relatively closed system.This paper has briefly introduced and analyzed the characteristics and main applications of knowledge structure in various fields from the aspects of literature,clinical medical cases,Traditional Chinese Medicine prescriptions and special disease specialties,and analyzes the knowledge representation technology and reasoning methods in the field of Traditional Chinese medicine.It is considered that the knowledge structure of TCM resources is different,and its fuzziness is the main bottleneck of the development of knowledge base;the representation learning method of word vector and case-based reasoning method have better application prospects in the field of TCM.It is helpful to extract deep level features and generalize the model from the modeling of syndrome element.In the future,corresponding standards and specifications should be formulated in a unified way,and a global view of the field should be built by sharing and co construction.At the same time,knowledge review and update mechanisms should be emphasized to improve the quality of knowledge.