中醫症狀知識本體的建構

Design and Implementation of a Traditional Chinese Medicine Symptom Ontology

陳南宏(Nan-Hong Chen)

指導教授 ： 林迺衛

國立中正大學/工學院/資訊工程學系/碩士(2015年)

摘要

中醫是世界古老醫學之一，幾千年來，積累了相當多的臨床經驗與文獻，形成了獨特的診療體系。知識本體為某一特定領域的模型，利於該領域知識的擷取及推論。一個知識本體包含該領域術語的定義，及術語之間的關係。藉由知識本體，資訊系統可以分享領域知識，並從事有系統的領域知識推論。因此，中醫知識本體的建構是中醫資訊化的基礎建設。
中醫症狀知識本體是中醫知識本體最基礎的一部分。中醫症狀知識本體的核心工作是收集及定義中醫症狀術語，這是一個非常浩大的工程，屬於中醫專業領域的重要工作。與此同時，中醫症狀知識本體的建構有賴於資訊專業領域的協助。因此，本論文是跨中醫及資訊專業領域的研究成果。在本論文建構的中醫症狀知識本體中，一個症狀可由兩個症狀核心屬性：症狀物件與症狀屬性，及兩個症狀附屬屬性：症狀程度與症狀期間所組成。
中醫症狀知識本體有許多應用。本論文利用中醫症狀知識本體建構一個中醫症狀詞庫，從事中醫症狀的標準化。一般中醫文獻中的症狀描述存在許多一義多詞或一詞多義的情況，中醫症狀知識本體中的症狀術語皆為一詞一義。透過中醫症狀知識本體，中醫症狀詞庫可以將多個一義多詞的同義症狀都對應到同一個一詞一義的症狀術語，也可以將一個一詞多義的症狀拆解成多個一詞一義的症狀術語。本論文也利用中醫症狀知識本體建構一個中醫辨證系統的雛型，從事中醫辨證的自動化。中醫症狀標準化是中醫辨證自動化的基礎，利於從事精確量化的中醫辨證。中醫症狀知識本體的建構必須經過長期的演化，才會漸趨正確與完整。希望本研究可以啟動一個正確與完整的中醫症狀知識本體的建構。

關鍵字

知識本體 ； 症狀標準化 ； 中醫症狀詞庫 ； 中醫辨證

並列摘要

Traditional Chinese medicine (TCM) is one of the ancient medicines on the world. Through thousands of years, it has evolved into a unique system for medical diagnosis and treatment, and a large volume of clinical experiences and literatures has been accumulated. An ontology is a model of a particular domain that facilitates the capturing and inference of domain knowledge. An ontology includes the definitions of a set of terminologies and the relationships among these terminologies. Based on the ontology, information systems can share domain knowledge and perform systematic inference of domain knowledge. Therefore, the design and implementation of a TCM ontology provides an infrastructure for applying information technology to TCM.
A TCM symptom ontology is the most basic component of a TCM ontology. The core task of a TCM symptom ontology is the collection and definition of symptom terminologies. This task is vast and depends on the domain knowledge of TCM. On the other hand, the design and implementation of a TCM symptom ontology depends on the domain knowledge of information engineering. This thesis is an interdisciplinary research achievement between TCM and information engineering. In our TCM symptom ontology, a symptom is defined by two primary attributes: the symptom object and the sympton property, and two auxiliary attributes: the sympton degree and the sympton duration.
A TCM symptom ontology has many applications. This thesis uses the TCM symptom ontology to develop a TCM symptom thesaurus to facilitate the standardization of TCM symptoms. There exist many symptom descriptions in the form of polysemy or synonyms in TCM literatures. All the symptom terminologies in the TCM symptom ontology have single variant. Based on the TCM symptom ontology, the TCM symptom thesaurus can map synonyms into the same symptom terminology, and divide a polysemy into several symptom terminologies. This thesis also uses the TCM symptom ontology to develop a prototypical TCM syndrome differentiation system to facilitate the automation of TCM syndrome differentiation. The standardization of TCM symptoms is the basis of the automation of TCM syndrome differentiation and facilitates the precise quantitative analysis of syndrome differentiation. The correctness and completeness of a TCM symptom ontology can only be achieved through a long period of evolution. We wish this research starts up the evolution of a correct and complete TCM symptom ontology.

並列關鍵字

ontology ； standardization of symptoms and signs ； traditional Chinese medicine symptom thesaurus ； TCM syndrome

參考文獻

[24] L.-Y. Lin, C.-H. Liu, Z.-Y. Chen, C. C. Yeh, N.-W. Lin and M-H. Yeh, "The research on the Standardization Process of Traditional Chinese Medicine Symptoms," Taiwan Journal of Clinical Chinese Medicine, vol. 18, no. 1, pp. 1-13, March 2012.

[4] 李培甯, 基於模糊理論之中醫辨證, 台灣: 國立中正大學資訊工程學系暨研究所, 2015.

[5] 陳玟伶, 奠基於知識本體的中醫症狀詞庫系統, 台灣: 國立中正大學資訊工程學系暨研究所, 2014.

[27] Y. Yamagata, H. Kou, K. Kozaki, R. Mizoguchi, T. Imai and K. Ohe, "Ontological Model of Abnormal States and its Application in the Medical Domain," in Proceedings of the 4th International Conference on Biomedical Ontology, 2013.

[1] 季紹良, 余明哲, 陳國樹 且 詹寬仁, 中醫診斷學, 台北: 東大國書股份有限公司, 2008.