- 學位論文
中醫藥古籍文獻資訊分析－關鍵字詞擷取段落屬性的方法
Information Analysis of Ancient Chinese Medicine Literatures Document-Keywords to Retrieve Text-Paragraph
摘要
傳統中醫藥發展已有多年的歷史，先人的醫藥使用經驗成為極為珍 貴的醫學智慧結晶，更成為現代醫學發展的重要參考文獻。資訊與電腦 的躍進，人們把中醫藥古籍數位資訊匯入至電腦中得到最佳的保存與快 速的查詢效果。但是，大量的中醫藥古籍文獻讓使用者造成資料上的混 亂，在散亂的中醫藥古籍文獻段落中找出使用者需要的資料，是本研究 的主要目的。 本研究的資料來源是利用《本草綱目》(明．李時珍)、《本草求真》 (清代．黃宮繡)、《神農本草經》(漢代．作者佚)文字資料三本「本草 類」中醫藥古籍文獻210 多萬字做為本論有研究文章的段落屬性分析。 並運用TF(Term Frequency)、IDF(Inverse Document Frequency)、 TF-IDF 的概念、分析中醫藥古籍文章內文屬性段落定義，並計算關鍵字 詞權重，得到關鍵字詞，進而建立知識庫把計算得到的關鍵字詞匯入， 利用「資料探勘」(Data Mining)方法處理中醫藥古籍文章中屬於描述 藥物【氣味】、【歸經】、【主治(功效)】欄位特徵段落擷取。 本研究結合專家學者的知識整合而成的知識庫，加上「資料探勘」 (Data Mining) 的資訊技術輔助應用，萃取特徵關鍵字詞文章屬性段 落，而達到上述目的，幫忙專家學者對於有效率性的閱讀中醫藥古籍， 進而提升中醫藥的相關研究。
並列摘要
The development of traditional Chinese medicine for many years the history of the ancestors of Chinese medicine as a very valuable experience in the use of medical wisdom, modern medicine has become an important reference. Leap forward in information and computer, it is the ancient Chinese medicine into the digital information to your computer and save the best results fast query. However, a large number of ancient Chinese medicine literature information allows users to cause confusion, in the scattered literature of the ancient Chinese medicine to identify the paragraphs of information the user needs is the main purpose of this study. Source of this study is the use of "Compendium of Materia Medica" (Ming. Li), "seeking truth and herbal medicine" (the Qing Dynasty. Embroidered yellow palace), "Shen Nong's Herbal Classic" (the Han Dynasty. Author Yi) 3 written materials "category Bencao "Ancient Chinese medicine literature as more than 210 million words of research in this article, paragraph attribute analysis. And the use of TF (Term Frequency), IDF (Inverse Document Frequency), TF-IDF concept, analysis of ancient Chinese medicine text articles attribute definition paragraphs, and calculate the weight of keywords, the keywords, thereby building a knowledge base to calculated by means of keyword terms, the use of "data mining" (Data Mining) approach to ancient Chinese medicine is described in the article odor drug 【vapor】, 【meridian】, 【effect】 paragraphs capture the characteristics of field . In this study, combined with the knowledge of experts and scholars from the knowledge base integration, coupled with "data mining" (Data Mining) applications of information technology-assisted extraction keywords article attributes the characteristics of the paragraphs, and achieve the above purpose, the help of experts and scholars for the efficient reading of ancient Chinese medicine, Chinese medicine to enhance research.