The Thesaurus-Based Construction Of Knowledge Base For TCM Fundamental Theory

Language knowledge base is an essential and fundamental resource in NLP system. It is widely applied in fields such as intelligent information processing, knowledge engineering and so on. Recent years, with the development of research for domain knowledge, more and more researchers have found the fact that common knowledge base does not meet the requirement of solving domain information problems. As a result, domain oriented knowledge base construction becomes a hot spot in the research of NLP. Chinese medicine domain knowledge, which has been accumulated for thousands of years, has very high research value. However, the knowledge mostly exists in old books, which is not convenient for inheritance and spreading. So, the construction of specific and formal concept model of sharing for domain oriented Chinese medicine knowledge base, which can also be called the computer readable basic theory knowledge base for Chinese medicine, has great significance.This paper mainly focuses on the knowledge organizations for the concept expressed in the natural language from the Glossary of terms in the field of the fundamental theory of TCM, and then construct the Knowledge Base which contains the knowledge that computer can understand and deal with. The main work consists of the following parts:(1) Collect the semantic resources needed for the KB construction in the field of the fundamental theory of TCM, and analyze the description of the term and the relation between the terms from the glossary of terms. According to terms of features, we propose an auxiliary method of the KB construction by drawing lessons from the construction method of the semantic KB of HowNet.(2) Propose a system of conceptual description for the KB in the field of the fundamental theory of TCM, including the primitive acquisition of the fundamental theory of TCM and the relation acquisition. The primitive acquisition first automatically established the primitives of candidate by using the simple terms and the word frequency of the word segmentation results, then build a meaning of the primitive table,which contains a total of 99 primitives, then transform the problem of relation acquisition into a classification problem on the basis of analysis of seven kinds of knowledge relation types, using semantic similarity calculation method to achieve.(3) Design and implement a knowledge base auxiliary building systems, which mainly divided into knowledge acquisition and knowledge retrieval two modules. Finally using this system builds a knowledge base that has been implemented on the conceptual description of 553 terms of the fundamental theory of TCM.