中藥的療效和其品質有著密切的關係，若能對中藥進行質量控制，進一步中藥的療效也能相對的穩定。由於中藥對環境的影響非常敏感，而傳統之鑑別技術因人為主觀意識濃厚，無法完全解決中藥的模糊性及複雜性，因此在中藥市場上充斥著正品、非正品及偽品，同名異物、同物異名的情況非常普遍。近來科學技術利用中藥指紋圖譜去表徵一個中藥，開啟中藥鑑別的新紀元。因此本研究以Possibility c-Means (PCM)為基礎，提出一架構平台，從中藥指紋圖譜抽取重要特徵，並從已知樣品訓練模糊分類器(fuzzy classifier)，以PCM 為基礎之模糊分類器將考慮檢品屬於不同品種類別的可能性(possibility)。藉以改善傳統鑑別技術之不足。本架構平台將提供具最高歸屬可能性的預測類別及依相關係數(correlation coefficient)計算檢品和預測類別間之相似度做監控上的參考。As we know, the quality of Chinese Medicine plays an important role in curative effect. The conventional identification techniques for Chinese Medicine use experiences of experts of Chinese Medicine to recognize Chinese Medicine. It is subjective and not suited to some types of Chinese Medicine. So, we make use of the concept of Possibility c-Means (PCM) to construct fuzzy classifier from the training samples. In order to improve the accurate matching rate, fuzzy classifier takes into account the possibility of test sample belong to any class. It is reasonable because the Chinese Medicine is sensitive to the environment and having the property of fuzziness.Our results of research provide possibilities of prediction classes and similarities between the unknown sample and prediction classes.
