文字探勘技術應用於中醫診斷腦中風之研究

近代中醫學家對腦中風的辨證論治，各有各的想法和所長。有的主張要辨陰虛或陽虛；有的主張以痰熱腑實為主；有的著重痰瘀相兼等等。而在中醫論述腦中風方面，不論以前各個朝代或是近代的中醫師在想法和方法，差異最大的就是在辯證和病因認定方面，所以本研究著重在這兩方面的研究，利用文字探勘技術，萃取出文章內的關鍵詞與其關聯架構，以達到以下的目的：第一、發現中藥辨證腦中風常見的方式和症候、用藥的關聯性，第二、以「腦中風之證型、症狀和用藥」這部分的資料，建立一個有效的文件分類指標，第三、萃取出腦中風的病因和關聯架構，期望可以做進一步的分析探討。研究結果顯示，本研究找出各個證型的關聯架構，並以SVM進行驗證，冀以整理出中醫在辦證腦中風上有益的方法與想法。可以發現更多隱含的資訊，進一步協助中醫在腦中風方面的研究。

關鍵字

腦中風；中醫；病因；文字探勘

並列摘要

For ages, upon the topic of treatments of Stroke, Chinese medicine experts have debating on the therapy and recognizing the causes of Stroke. This research will focus on these two debates, with using Text Mining to find out the keyword in the article and the treatments of each syndrome in order to fulfill: First, to find out the common Chinese medicine therapy upon Stroke and the connection between the symptom and the treatment; Second, to use the information about ”the therapy, symptom and the treatment of Stroke” to establish an effective classification index; Thirdly, to find out the disease causation of Stroke, the treatments of each syndrome, and hoping to do the further analysis. The results indicate that: this research find out the treatments of each syndrome, testified it by SVM program and hoping to get further information to assist Chinese medical experts on the fields of Stroke.